Patent Application
20080160558
A process for generating sequence-specific proteases by screening-based directed evolution is disclosed. The use of the process provides proteases recognizing and cleaving user-definable amino-acid sequences with high sequence-specificity. Proteases obtainable by the process can be used in a variety of medical, diagnostic and industrial applications.
Proteolytic enzymes or proteases are a class of enzymes which has an outstanding position among the different enzymes, since the reaction catalyzed by proteases is the cleavage of peptide bonds in other proteins. Proteases are not only very common enzymes in nature, but belong to the most important enzymes for medical and industrial use. Of the total worldwide sales of enzymes, which is estimated to be more than USD 1 billion per year, proteases account for approximately 60%. Based on the functional group present at the active site, proteases are classified into four groups, i.e., serine proteases (EC 3.4.21), cysteine proteases (EC 3.4.22), aspartic proteases (EC 3.4.23), and metalloproteases (EC 3.4.24). Classification into one of the four groups is typically done by experimental determination of sensitivity towards different types of protease inhibitors. Furthermore, proteases of the four groups differ in their biochemical properties. For example, serine proteases are sensitive to inhibitors 3,4-DCI, DFP, PMSF and TLCK, and have a pH optimum between pH 7 and 11. Aspartic proteases are inhibited by pepstatin, DAN and EPNP, and predominantly have a pH optimum between pH 3 and 4. Cysteine proteases are sensitive to sulfhydryl inhibitors such as PCMB, and besides a few exceptions, have neutral pH optima. Metalloproteases are characterized by the requirement of a divalent metal ion for their activity. Therefore, metalloproteases are inhibited by chelating agents such as EDTA, and have neutral or alkaline pH optima. Among these four groups, further classification is usually done on the basis of structural similarities.
Besides such a combined biochemical and structural classification, proteases can be grouped according to their substrate spectrum. The two most general groups to be distinguished are exoproteases and endoproteases. Exoproteases only cleave peptide bonds at the very end of an peptide, whereas endoproteases catalyze the cleavage of bonds anywhere in a peptide strand. The specificity of proteases, i.e their ability to recognize and hydrolyze specifically certain peptide substrates while others remain uncleaved, can be expressed qualitatively and quantitatively. Qualitative specificity refers to the kind of amino acid residues that are accepted by a protease at certain positions of the peptide substrate. For example, trypsin and the tissue-type plasminogen activator are related with respect to their qualitative specificity, since both of them require at the position P1 an arginine or a similar residue (nomenclature of peptide substrate positions according to the nomenclature of Schlechter & Berger (Biochem. Biophys. Res. Commun. 27 (1967) 157-162). On the other hand, quantitative specificity refers to the relative number of peptide substrates that are accepted as substrates. The quantitative specificity can be expressed by the term
S=−log(Q),
where Q is the ratio of all accepted peptide substrates versus all possible peptide substrates. Quantitative specificities of several proteases are shown exemplarily in Table I. The calculation of quantitative specificities is based on the twenty naturally occurring amino acids, and on the assumption that all combinations of these twenty amino acids are feasible. Consequently, proteases that accept only a small portion of all possible peptides have a high specificity, whereas the specificity of proteases that, as an extreme, cleave any peptide substrate would theoretically be zero.
The quantitative specificity of proteases varies over a wide range. There are very unspecific proteases known, such as papain which cleaves all polypeptides that contain a phenylalanine, a valine or an leucine residue (s=0.82), or trypsin which cleaves all polypeptides that contain an arginine or a lysine residue (s=1.0). On the other hand, there are highly specific proteases known, such as the tissue-type plasminogen activator (t-PA) which cleaves plasminogen only at a single specific sequence (s=9.11). Proteases with high substrate specificity play an important role in the regulation of protein functions in living organisms. The specific cleavage of polypeptide substrates, for example, activates precursor proteins or deactivates active proteins or enzymes, thereby regulating their functions. Several proteases with high substrate specificities are used in medical applications. Pharmaceutical examples for activation or deactivation by cleavage of specific polypeptide substrates are the application of t-PA in acute cardiac infarction which activates plasminogen to resolve fibrin clots, or the application of Ancrod in stroke which deactivates fibrinogen, thereby decreasing blood viscosity and enhancing its transport capacity. While t-PA is a human protease with an activity necessary in human blood regulation, Ancrod is a non-human protease. It was isolated from the viper Agkistrodon rhodostoma, and comprises the main ingredient of the snake's poison. Therefore, there exist a few non-human proteases with therapeutic applicability. Their identification, however, is usually highly incidental.
The treatment of diseases by administering drugs is typically based on a molecular mechanism initiated by the drug that activates or inactivates a specific protein function in the patient's body, be it an endogenous protein or a protein of an infecting microbe or virus. While the action of chemical drugs on these targets is still difficult to understand or to predict, protein drugs are able to specifically recognize these target proteins among millions of other proteins. Prominent examples of proteins that have the intrinsic possibility to recognize other proteins are antibodies, receptors, and proteases. Although there are a huge number of potential target proteins, only very few proteases are available today to address these target proteins. Due to their proteolytic activity, proteases are particularly suited for the inactivation or activation of protein targets. When considering human proteins only, the number of potential target proteins is yet enormous. It is estimated that the human genome comprises between 30,000 and 100,000 genes, each of which encodes a different protein. Many of these proteins are involved in human diseases and are therefore potential pharmaceutical targets Proteases recognizing and cleaving these target proteins with a high specificity are consequently of high value as potential drugs. The medical application of such proteases, however, is restricted by their occurrence. For example, there are theoretically 25 billion different possibilities for a specificity of s=10.4 (corresponding to the specific recognition of a unique sequence of eight amino acid residues). It is highly unlikely to find such a protease with one particular qualitative specificity by screening natural isolates.
Selection systems for proteases of known specificity are known in the art, for instance, from Smith et al., Proc. Natl. Acad. Sci USA, Vol. 88 (1991). As exemplified, the system comprises the yeast transcription factor GAL4 as the selectable marker, a defined and cleavable target sequence inserted into GAL4 in conjunction with the TEV protease. The cleavage separates the DNA binding domain from the transcription activation domain and therewith renders the transcription factor inactive. The phenotypical inability of the resulting cells to metabolize galactose can be detected by a calorimetric assay or by the selection on the suicide substrate 2-deoxygalactose.
Further, selection may be performed by the use of peptide substrates with modifications as, for example, fluorogenic moieties based on groups as ACC, previously described by Harris et al. (US 2002/022243).
Laboratory techniques to generate proteolytic enzymes with altered sequence specificities are in principle known. They can be classified by their expression and selection systems. (Genetic selection means to produce a protease within an organism which protease is able to cleave a precursor protein which in turn results in an alteration of the growth behavior of the producing organism. From a population of organisms with different proteases those having an altered growth behavior can be selected. This principle was reported by Davis et al. (U.S. Pat. No. 5,258,289, WO 96/21009). The production of a phage system is dependent on the cleavage of a phage protein which only can be activated in the presence of a proteolytic enzyme or antibody which is able to cleave the phage protein. Selected proteolytic enzymes or antibodies would have the ability to cleave an amino acid sequence for activation of phage production. Furthermore, there is no control of the specificity of the proteases that are selected. The system does not select for proteases with low activities for other peptides than the used peptide substrate. Additionally, this system does not allow a precise characterization of the kinetic constants of the selected proteases (kcat, KM). Several other systems with intracellular protease expression are reported but they all suffer from the disadvantages mentioned above. Some of them use a genetic reporter system which allows a selection by screening instead of a genetic selection, but also cannot overcome the intrinsic insufficiency of the intracellular characterization of proteases.
A system to generate proteolytic enzymes with altered sequence specificities with membrane-bound proteases is reported. Iverson et al. (WO 98/49286) describe an expression system for a membrane-bound protease which is displayed on the surface of cells. An essential element of the experimental design is that the catalytic reaction has to be performed at the cell surface, i.e., the substrates and products must remain associated with the bacterium expressing the enzyme at the surface. This restriction limits the generation of proteolytic enzymes with altered sequence specificities and does not allow a precise characterization of the kinetic constants of the selected proteases (kcat, KM). Furthermore, the method does not allow the control of the position at which the peptide is cleaved. Additionally, positively identified proteases will have the ability to cleave a certain amino acid (aa) sequence but they also may cleave many other aa sequences. Therefore, there is no control of the specificity of the proteases that are selected.
A system to generate proteolytic enzymes with altered sequence specificities with self-secreting proteases is also known. Duff et al. (WO 98/11237) describe an expression system for a self-secreting protease. An essential element of the experimental design is that the catalytic reaction acts on the protease itself by an autoproteolytic processing of the membrane-bound precursor molecule to release the matured protease from the cellular membrane into the extracellular environment. Therefore, a fusion protein must be constructed where the target peptide sequence replaces the natural cleavage site for autoproteolysis. Limitations of such a system are that positively identified proteases will have the ability to cleave a certain aa sequence but they also may cleave many other peptide sequences. Therefore, high substrate specificity cannot be achieved with such an approach. Additionally, such a system is not able to control that selected proteases cleave at a specific position in a defined aa sequence and it does not allow a precise characterization of the kinetic constants of the selected proteases (kcat, KM).
Broad et al. (WO 99/11801) disclose a heterologous cell system suitable for the alteration of the specificity of proteases. The system comprises a transcription factor precursor wherein the transcription factor is linked to a membrane anchoring domain via a protease cleavage site. The cleavage at the protease cleavage site by a protease releases the transcription factor, which in turn initiates the expression of a target gene being under the control of the respective promotor. The experimental design of alteration of the specificity consists in the insertion of protease cleavage sites with modified sequences and the subjection of the protease to mutagenesis. New proteases obtained may be able to recognize the modified sequence, the effect of which is monitored by the expression of the target gene. Such a system does also not allow a precise control of biochemical properties of the selected proteases.
Most of these approaches apply methods of directed evolution for the generation of proteolytic enzymes with altered sequence specificities. Several different mutation and recombination methods to generate genetic libraries are reported and described elsewhere. All the different methods suffer from their lack of precise selection of positive protease variants from large libraries. First, these methods are not able to distinguish between single and multi turn-overs of peptide substrates which is necessary in order to prevent the selection of low kcat variants. Secondly, it is not possible to trigger enzyme and substrate concentration to select protease variants for lower KM. Third, none of these systems allows the selection of a protease with an increased activity on the desired peptide substrate whereby the activity on the original peptide substrate decreases.
Methods which fulfill the above mentioned three selection criteria (kcat, KM and substrate specificity) for generating proteolytic enzymes with high sequence-specificity applying screening-based directed evolution have heretofore not been available.
Thus, the technical problem underlying the present invention is to provide a method for generating new proteases with user-defined substrate specificities by applying directed evolution. In particular, the invention is directed to a method for the evolution of novel proteases towards selective recognition and cleavage of specific amino-acid sequences only. This technical problem has been solved by the embodiments of the invention specified below and In the appended claims. The present invention is thus directed to
The identification and selection of proteases that have evolved towards the target specificity is done by screening for catalytic activities on different peptide substrates, either by screening for increased affinity, or by using two substrates in comparison, or by using unspecific peptides as competitors, or by using intermediate peptide substrates.
The following detailed description will disclose the preferred features, advantages and the utility of the present invention.
The following figures are provided in order to explain further the present invention in supplement to the detailed description:
In the framework of this invention the following terms and definitions are used
The term “protease” means any protein molecule acting in the hydrolysis of peptide bonds. It includes naturally-occurring proteolytic enzymes, as well as variants thereof obtained by site-directed or random mutagenesis or any other protein engineering method, any fragment of an proteolytic enzyme, or any molecular complex or fusion protein comprising one of the aforementioned proteins. A “chimera of proteases” means a fusion protein out of two or more fragments derived from different parent proteases.
The term “substrate” or “peptide substrate” means any peptide, oligopeptide, or protein molecule of any amino acid composition, sequence or length, that contains a peptide bond that can be hydrolyzed catalytically by a protease. The peptide bond that is hydrolyzed is referred to as the “cleavage site”. Numbering of positions in the substrate is done according to the system introduced by Schlechter & Berger (Biochem. Biophys. Res. Commun. 27 (1967) 157-162). Amino acid residues adjacent N-terminal to the cleavage site are numbered P1, P2, P3, etc., whereas residues adjacent C-terminal to the cleavage site are numbered P1′, P2′, P3′, etc.
The term “specificity” means the ability of a protease to recognize and hydrolyze selectively certain peptide substrates while others remain uncleaved. Specificity can be expressed qualitatively and quantitatively. “Qualitative specificity” refers to the kind of amino acid residues that are accepted by a protease at certain positions of the peptide substrate. “Quantitative specificity” refers to the number of peptide substrates that are accepted as substrates. Quantitative specificity can be expressed by the term s, which is the negative logarithm of the number of all accepted peptide substrates divided by the number of all possible peptide substrates. Proteases that accept only a small portion of all possible peptide substrates have a “high specificity” (s>>1). Proteases that accept almost any peptide substrate have a “low specificity”. Proteases with very low specificity (s≦1) are also referred to as “unspecific proteases”.
The term “first protease” describes any protease used in step (a) of this invention as the starting point in order to generate populations of protease variants that are related to this first protease. The term “first substrate” or “first peptide substrate” describes a substrate that is recognized and hydrolyzed by the first protease. The term “first specificity” describes the qualitative and quantitative specificity of the first protease.
The term “evolved protease” describes any protease that is generated by use of the method of the invention. The term “target substrate” or “target peptide substrate” describes a substrate that is recognized and hydrolyzed by the evolved protease. The term “target specificity” describes the qualitative and quantitative specificity of the evolved protease that is to be generated by use of the method of the invention. Thus, the target specificity defines the specificity of the evolved protease for the target peptide substrate while other substrates are not or very weakly recognized and hydrolyzed.
The term “intermediate” or “intermediate substrate” describes any substrate that has an intermediate character between two other substrate. The intermediate character can base on the amino acid composition, the amino acid sequence, the properties of the amino acid residues contained in the substrates, or a combination of these characteristics.
Catalytic properties of proteases are expressed using the kinetic parameters “KM” or “Michaelis Menten constant”, “kcat” or “catalytic rate constant”, and “kcat/KM” or “catalytic efficiency”, according to the definitions of Michaelis and Menten (Fersht, A., Enzyme Structure and Mechanism, W. H. Freeman and Company, New York, 1995). The term “catalytic activity” describes the rate of conversion of the substrate under defined conditions.
Amino acids are abbreviated according to the following Table II either in one- or in three-letter code.
As set forth above, the present invention is directed to a method for generating sequence-specific proteases with a target substrate specificity by applying principles of molecular evolution According to the invention, this is achieved by providing a population of proteases being related to each other, as well as a peptide substrate that resembles the target substrate, and selecting one or more protease variants from the population of proteases with respect to their specificity for the provided substrate. The selection is done under conditions that allow identification of proteases that recognize and hydrolyze the target sequence preferably.
In particular, embodiment (1) of the invention relates to a method for generating sequence-specific proteases with target substrate specificities, wherein the following steps are carried out:
When repeating steps (a) to (c), the one or more proteases selected in step (c) of one cycle are used as the one or more first proteases in step (a) of the next cycle.
In one alternative of the invention, the one or more first proteases serving as starting points in step (a) of the method have a high sequence specificity which is maintained high during the directed evolution towards the target specificity.
In another alternative of the method, the one or more first proteases serving as starting points in step (a) of the method have a low sequence specificity, which is increased during the directed evolution towards the target specificity.
The steps (a) to (c) of the above method are carried out for at least one cycle. Preferably, however, these steps are carried out for several cycles, with each one or more protease variants selected in one cycle being the origin of the population of protease variants in the next cycle. Preferably, more than one and less than hundred, more preferably more than two and less than fifty, particularly preferably more than three and less than twenty, especially preferably more than four and less than ten, and most preferably five cycles of steps (a) to (c) are carried out until one or more protease variants with the target substrate specificity are identified.
The invention applies evolutionary means as described in very detail in WO9218645 with that document being incorporated in its entirety for all purposes.
For an overview on the application of evolutionary principles to molecular biotechnology, which is usually referred to as “directed evolution” or “evolutionary biotechnology”, see the review by Koltermann & Kettling (Biophys. Chem. 66 (1997) 159-177).
Part of the invention is the provision of populations of protease variants wherein each variant is related to one or more first proteases. In principle, there can be a large number of these first proteases, all together being the origin for the first cycle of the method. It is preferred, however, that these first proteases comprise fifty or less different proteases, more preferably ten or less different proteases, especially preferably two or less different proteases. Most preferably, only one first protease is employed.
According to the invention, any protease can be used as first protease. Preferably, an endoprotease is used as first protease. It is preferred that the protease belongs to the group of proteases consisting of Serine proteases (EC 3.4.21), Cysteine proteases (EC 3.4.22), Aspartic proteases (EC 3.4.23), and Metalloproteases (EC 3.4.24). First proteases are characterized by their ability to recognize and hydrolyze peptide substrates with a certain qualitative and quantitative specificity First proteases can have a specificity in the same range as the specificity of the protease that is to be generated. Examples for proteases with relatively high specificities are TEV protease, HIV-1 protease, BAR1 protease, Factor Xa, Thrombin, tissue-type plasminogen activator, Kex2 protease, TVMV-protease, RSV protease, MuLV protease, MPMV protease, MMTV protease, BLV protease, EIAV protease, SIVmac protease. Alternatively, the first proteases have a lower specificity than the specificity of the protease that is to be generated. As an extreme example of the latter, proteases with very low sequence specificity are employed, for example proteases such as Papain, Trypsin, Chymotrypsin, Subtilisin, SET (trypsin-like serine protease from Streptomyces erythraeus), Elastase, Cathepsin G or Chymase.
A particularly suitable protease is sp|P12630| BAR1 protease (BAR1_YEAST Barrierpepsin precursor (EC 3.4.23.35) (extracellular “barrier” protein) (BAR proteinase) of S. cerevisiae (see SEQ ID NO:8).
The provision of populations of proteases is essentially done as described in WO9218645. According to the invention, genes encoding protease variants are ligated into a suitable expression vector by standard molecular cloning techniques (Sambrook, J. F; Fritsch, E. F.; Maniatis, T.; Cold Spring Harbor Laboratory Press, Second Edition, 1989, New York). The vector is introduced in a suitable expression host cell, which expresses the corresponding protease variant. Particularly suitable expression hosts are bacterial expression hosts such as Escherichia coli or Bacillus subtilis, or yeast expression hosts such as Saccharomyces cerevisae or Pichia pastoris, or mammalian expression hosts such as Chinese Hamster Ovary (CHO) or Baby Hamster Kidney (BHK) cell lines, or viral expression systems such as the Baculovirus system. Alternatively, systems for in vitro protein expression can be used.
In a preferred embodiment of the invention, the genes are ligated into the expression vector behind a suitable signal sequence that leads to secretion of the protease variants into the extracellular space, thereby allowing direct detection of protease activity in the cell supernatant. Particularly suitable signal sequences for Escherichia coli are HlyA, for Bacillus subtilis AprE, NprB, Mpr, AmyA, AmyE, Blac, SacB, and for S. cerevisiae Bar1, Suc2, Matα, Inu1A, Ggplp.
In another preferred embodiment of the invention, the protease variants are expressed intracellularly and the peptide substrates are expressed also intracellularly. Preferably, this is done essentially as described in WO 0212543, using a fusion peptide substrate comprising two auto-fluorescent proteins linked by the substrate amino-acid sequence.
In another preferred embodiment of the invention, the protease variants are expressed intracellularly, or secreted into the periplasmatic space using signal sequences such as DsbA, PhoA, PelB, OmpA, OmpT or gIII for Escherichia coli, followed by permeabilisation or lysis step to release the protease variants into the supernatant. The destruction of the membrane barrier can be forced by the use of mechanical means such as ultrasonic, French press, or the use of membrane-digesting enzymes such as lysozyme.
As a further alternative, the genes encoding the protease variants are expressed cell-free by the use of a suitable cell-free expression system. In a particularly preferred embodiment, the S30 extract from Escherichia coli cells is used for this purpose as described by Lesly et al. (Methods in Molecular Biology 37 (1995) 265-278).
The relatedness to the one or more first proteases can be achieved by several procedures. For example, the genes encoding the one or more first proteases are modified by methods for random nucleic acid mutagenesis. In a preferred embodiment of the invention, random mutagenesis is achieved by the use of a polymerase as described in WO 9218645. According to this embodiment, the one or more genes encoding the one or more first proteases are amplified by the use of a polymerase with a high error rate, or under conditions that increase the rate of misincorporations, thereby leading to a population of genes wherein each gene encodes a protease that is related to the one or more first proteases. For example the method according to Cadwell, R. C and Joyce, G. F. can be employed (PCR Methods Appl. 2 (1992) 28-33). Other methods for random mutagenesis that can be employed make use of mutator strains, UV-radiation or chemical mutagens. Most preferably, errors are introduced into the gene at or near but below the error threshold as described in WO 9218645.
In another preferred embodiment of the invention, certain parts of the gene encoding the protease variants are randomized completely with respect to the amino-acid sequence, and are re-introduced into the gene as an oligonucleotide cassette. This technique is usually referred to as cassette mutagenesis (Oliphant, A. R. et al., Gene 44 (1986) 177-183; Horwitz, M. S., et al. Genome 31 (1989) 112-117). In a particularly preferred embodiment of the invention, the part of the gene that encodes amino acid residues that are essential for recognition of the substrate is randomized via cassette mutagenesis. These residues can be identified from structural studies. In particular, residues comprising parts of the substrate binding pocket are targeted by cassette mutagenesis. Alternatively, substituting each amino acid residue with an alanine, and analyzing whether there is an effect on the catalytic activity can identify such residues. As a further alternative, these residues can be identified by first introducing random mutations into the gene, screening for an effect on specificity, affinity, or catalytic activity, and determining afterwards the position of mutations in variants that represent altered specificity, affinity or altered catalytic activity. As an extreme of this approach, the completely randomized sequence can has the length of one nucleotide only. This approach is typically referred to as site saturation mutagenesis.
In another preferred embodiment of the invention, nucleic acid sequences are randomly introduced into or deleted from the one or more first protease genes in order to provide a population of proteases. This approach is referred to as insertion and/or deletion mutagenesis. For insertion mutagenesis, random sequences of defined or random length are introduced randomly into a gene. As an example, the method described by Hallet et al (Nucleic Acids Res. 1997, vol. 25, p. 1866ff) can be used to introduce a random 15 nt sequence randomly into a gene. Alternatively, defined sequences, for example a sequence encoding a specific protein secondary structure motif, can he inserted randomly into a gene. Alternatively, random sequences of defined or random length can be inserted at specific sites into a gene. This can be done using restriction sites or by oligonucleotide overlap extension methods such as the method described by Horton (Gene 1989, vol. 77, p. 61ff). For deletion mutagenesis, sequences of defined or random length are deleted randomly from a gene. In a particular embodiment of the invention, deletion and insertion mutagenesis are combined so that insertions at one site can potentially be combined, and thereby possibly compensated, by deletion at another site.
In a further preferred embodiment of the invention, methods for homologous in-vitro recombination are used for the provision of protease populations. Examples of methods that can be applied are the Recombination Chain Reaction (RCR) according to WO 0134835, the DNA-Shuffling method according to WO 9522625, the Staggered Extension method according to WO 9842728, or the Random Priming recombination according to WO9842728. Furthermore, also methods for non-homologous recombination such as the Itchy method can be applied (Ostermeier, M. et al., Nature Biotechnology 17 (1999) 1205-1209). All of the references mentioned above are hereby incorporated by reference in its entirety for all purposes.
In further embodiments of the invention, the above-mentioned methods are combined with each other. In a particularly preferred embodiment, the Recombination Chain Reaction is combined with random mutagenesis such as error-prone PCR according to Cadwell, R. C and Joyce, G. F. (PCR Methods Appl. 2 (1992) 28-33) in order to de-couple mutations selected in the round before and to introduce simultaneously a defined number of new random mutations into the population.
The coupling of protease genotype and phenotype is achieved by use of sample carriers that enable compartmentation of samples, and the distribution of genotypes into sample carriers is done at a multiplicity per compartment that allows sufficient differentiation of phenotypes.
The one or more first proteases that serve as the starting point of the method either have a specificity which is in the range of the target specificity that is to be generated by the method, or have a lower specificity than the target specificity. Accordingly, the method of the invention is either performed under conditions that maintain the specificity quantitatively and alters it qualitatively (Alternative A), or the method of the invention is performed under conditions that maintains the specificity qualitatively and increases it quantitatively (Alternative B). Moreover, both approaches can be combined. These three principle alternatives are shown schematically in
In a preferred embodiment of the invention corresponding to alternative A, the one or more first proteases have a first specificity that is quantitatively in the range of the target specificity, but qualitatively distinct from the target specificity. Proteases having the target substrate specificity are achieved using the method of the invention by selecting protease variants under conditions that allow identification of proteases that recognize and cleave the target sequence preferably.
In another preferred embodiment of the invention corresponding to alternative B, the specificity of the one or more first proteases is quantitatively lower when compared to the target specificity. This means that they accept and hydrolyze a larger number of peptide substrates. This low first specificity is subsequently increased by the method of the invention until it is in the range of the target specificity. As a preferred variant of this embodiment, the first specificity is qualitatively related to the target specificity. Thus, the large number of peptide substrates that is accepted and hydrolyzed includes the target substrate already. Accordingly, amino acid residues that are essential in the first substrate remain essential residues in the target substrate. Then, proteases having the target substrate specificity are achieved using the method of the invention by selecting protease variants under conditions that allow identification of proteases that recognize and cleave the target sequence preferably.
Another part of the invention is the provision of peptide substrates that resemble the target substrate, and the use of these substrates for screening of protease variants with respect to their catalytic activity.
In a preferred embodiment of the invention, suitable peptide substrates are synthesized via the solid phase peptide synthesis approach of Merrifield et al (Nature. 207 (1965) 522-523). These peptide substrates are then incubated for a certain time in a sample buffer containing the protease variant to be tested. The hydrolysis of the peptide is then analyzed by a suitable method. For example, the amount of fragmented peptides can be analyzed by chromatography. In particular, peptide fragments are analyzed advantageously on a reversed phase HPLC system. Alternatively, the peptide substrate is modified in any way to enable the analysis of peptide hydrolysis. In particular, the peptide substrate may carry functional groups that enable the detection of the hydrolysis of the substrate. Such functional groups include, but are not limited to, the following:
With respect to the above mentioned functional groups, a chemical group can be attached to the peptide that alters its properties when the peptide is hydrolyzed. For example, a para-nitrophenyl group can be used for this purpose. As another example, one or more fluorophores and/or a quencher molecule are attached to the peptide, and the amount of fragmented peptide is analysed by measuring a difference in the fluorescence of the fluorophors. For example, two fluorophores that are suited to form a FRET (fluorescence resonance energy transfer) pair are attached to the peptide at opposite ends, and the hydrolysis of the peptide is measured by a decrease in the energy transfer between the two fluorophors. For example, Rhodamine Green (Molecular Probes Inc., Oregon, USA) and Tetramethylrhodamine (Molecular Probes Inc., Oregon, USA) can be used as fluorophors that are suited to form such a FRET pair.
In a particularly preferred embodiment of the invention, two fluorophores that do not form a substantial FRET pair are attached to opposite ends of synthetic peptide substrates. As an example, Rhodamine Green (Molecular Probes Inc., Oregon, USA) and Cy-5 (Amersham Biosciences Europe GmbH, Freiburg) can be used for this purpose, and covalent attachment of the dye can be achieved via a succinimidyl ester linkage to a primary amino group of the peptide. Hydrolysis of these peptides is preferably analysed by means of confocal fluorescence spectroscopy according to patent applications WO 9416313 and WO9613744, which are hereby incorporated by reference in their entirety for all purposes. Due to the high sensitivity of confocal fluorescence spectroscopy, substrates are used in concentrations below one micromolar, more preferably below hundred nanomolar, and most preferably below ten nanomolar. Therefore, screening according to this embodiment is done substantially below the KM of typical proteases.
In another particularly preferred embodiment of the invention, fusion proteins comprising a first autofluorescent protein, a peptide, and a second autofluorescent protein are used as peptide substrates. According to WO0212543, which is hereby incorporated by reference in its entirety for all purposes, autofluorescent include the Green Fluorescent Protein GFP and its mutants, as well as dsRED and its mutants. Fusion proteins can be produced by expression of a suitable fusion gene in E. coli, lysis of cells and purification of the fusion protein by standard methods such as ion exchange chromatography or affinity chromatography.
It is an essential part of the invention that proteases with the target substrate specificity are generated by selecting protease variants under conditions that allow identification of proteases that recognize and cleave the target sequence preferably. This selection can be achieved according to the different aspects of the invention as outlined below.
In a first aspect of the invention, proteases that recognize and cleave the target sequence preferably are identified by screening for proteases with a high affinity for the target substrate sequence. High affinity corresponds to a low KM which is selected by screening at target substrate concentrations substantially below the KM of the first protease. This aspect is referred to as the “affinity approach”.
In a preferred embodiment of this aspect of the invention, the peptide substrate provided in step (b) is linked to one or more fluorophores that enable the detection of the hydrolysis of the peptide substrate at concentrations below 10 μM, preferably below 1 μM, more preferably below 100 nM, and most preferably below 10 nM.
In a second aspect of the invention, proteases that recognize and cleave the target sequence only are identified by providing two or more peptide substrates in step (b) and by screening for activity on these two or more peptide substrates in comparison. This aspect is referred to as the “comparison approach”.
In a preferred embodiment of this aspect of the invention, the two or more peptide substrates provided in step (b) are linked to different marker molecules, thereby enabling the detection of the cleavage of the two or more peptide substrates consecutively or in parallel. In a particularly preferred embodiment of the invention, two peptide substrates are provided in step (b), one peptide substrate having an amino-acid sequence identical to or resembling the first peptide substrate thereby enabling to monitor the original activity of the first proteases, and the other peptide substrate having an amino-acid sequence identical to or resembling the target substrate sequence thereby enabling to monitor the activity on the target substrate. In an especially preferred embodiment of the invention, these two peptide substrates are linked to fluorescent marker molecules, and the fluorescent properties of the two peptide substrates are sufficiently different in order to distinguish both activities when measured consecutively or in parallel. For example, a fusion protein comprising a first autofluorescent protein, a peptide, and a second autofluorescent protein according to patent application WO 0212543 can be used for this purpose. Alternatively, fluorophores such as rhodamines are linked chemically to the peptide substrates.
In a third aspect of the invention, proteases that recognize and cleave the target sequence preferably are identified by providing in step (b) one or more peptide substrates resembling the target peptide together with competing peptide substrates in high excess. Screening with respect to activity on the substrates resembling the target substrate is then done in the presence of the competing substrates. Proteases having a specificity which corresponds qualitatively to the target specificity, but having only a low quantitative specificity are identified as negative samples in such a screen. Whereas proteases having a specificity which corresponds qualitatively and quantitatively to the target specificity are identified positively. This aspect is referred to as the “competitor approach”.
In a preferred embodiment of this aspect of the invention, the one or more peptide substrates resembling the target substrate are linked to marker molecules, thereby enabling the detection of their hydrolysis, whereas the competing peptide substrates do not carry marker molecules. The competing peptide substrates have an amino-acid sequence identical to or resembling the first peptide substrate, or have random amino-acid sequences, thereby acting as competitive inhibitors for the hydrolysis of the marker-carrying peptide substrates.
In a fourth aspect of the invention, proteases that recognize and cleave the target sequence preferably are identified by using intermediate substrates for evolving the protease towards the target substrate specificity. This aspect is hereinafter also referred to as the “intermediate approach”. In a first variant of this aspect of the invention, this is achieved by providing in different cycles different peptide substrates, whereby each peptide substrate has an intermediate character with regard to the cycle before and the target peptide substrate. According to this variant, proteases are evolved gradually toward the target specificity.
More generally, a first variant of this aspect of the invention is directed to a method for generating sequence-specific proteases with a target substrate specificity, wherein the following steps are carried out:
In this first variant of this aspect of the invention, evolution of protease specificity is directed via consecutive selection on a certain number of intermediate peptide substrates, whereby every peptide substrate resembles more and more the target peptide sequence, This approach is based on the finding that proteases which accept related substrates are usually also related to each other. Relatedness of proteases in the context of this invention is a measure for the homology in the amino acid sequences of two or more enzymes Moreover, this approach is based on the surprising discovery, that distinguishable subsites in a protease active site can be evolved separately, and that their molecular structure can be attributed to different residues of a peptide substrate (Schlechter & Berger, Biochem. Biophys, Res. Commun. 27 (1967) 157-162).
Intermediate substrates can be realized by substituting amino acid residues at one or more positions from the first peptide sequence with amino acid residues at the same positions from the target peptide sequence. Such intermediates are referred to as “amino acid composition intermediates”. Additionally, an intermediate peptide substrate can include one or more amino acid residues at one ore more positions which are neither the residues of the first peptide sequence nor the residues of the target peptide sequence at that position, but are amino acid residues with an intermediate character with respect to the residues in the first and the target substrate. Such intermediates are referred to as “amino acid property intermediates”. The intermediate character of this kind of intermediates can be based on one or more physical and chemical parameters, which include, but are not limited to, the surface of the residue, its volume, the isoelectric point, the side chain pKa, the polarity, the ability to form hydrogen bonds or the hydrophobicity. In the following table, the twenty naturally occurring amino acid residues are classified according to these parameters.
aChothia, C., J. Mol. Biol., 105 (1975) 1-14;
bZamyatin, A. A., Prog. Biophys. Mol. Biol., 24 (1972) 107-123;
cTanford, C., Adv. Prot. Chem., 17 (1962) 69-165;
dcharge at physiological pH;
eBlack, S. D, Mould, D. R, Anal. Biochem., 193 (1991) 72-82.
If, for example, the first substrate were ALY and the target substrate were NRF, intermediate substrates with regard to the amino acid composition would be, for example, ALF, NRY, or ARF (Modifications with regard to the first substrate are indicated). An intermediate substrate with regard to amino acid properties would be, for example, AQF, with the glutamine residue resembling the original leucine in the sense that it is uncharged, but resembling more the arginine residue of the target substrate with respect to hydrophobicity and its capacity to form hydrogen bonds. A further example for this approach would be SLY where S resembles A with respect to volume and surface but is more similar to the target N in terms of hydrophobicity and hydrogen bonding.
Furthermore, the number of consecutive peptide substrates to be used depends on the relatedness of the first peptide sequence and the target peptide sequence as well as the quantitative specificity of the one or more first proteases. The more unrelated the first peptide sequence and the target peptide sequence are, and the higher the specificity of the one or more first proteases is, the more consecutive intermediate peptide substrates are required.
In a second variant of this aspect of the invention, different proteases that have specificity for different intermediates are selected in parallel in a first step of the method. In a second step, proteases which have the target specificity are then selected from a population containing randomly recombined chimeras of the proteases selected in the first step. Preferably, the recombination of different proteases selected in parallel is achieved by employing an in-vitro homologous recombination technique, such as the Recombination Chain Reaction described in patent application WO 0134835. Both forms of intermediates can be used for this variant. However, amino acid composition intermediates are preferably employed.
The different intermediates employed in the first step of this variant are preferably chosen in a way, that the sum of all modifications introduced into these intermediates equals or resembles the characteristics of the target substrate. As an example, when the first substrate were ALY and the target substrate were NRF, suitable amino acid composition intermediates for this embodiment would be NLY, ARY, and ALF. Proteases having specificity for these three substrates would then be randomly recombined and screened for specificity towards the target substrate of this example NRF. Other examples, including intermediates who more than one modification are to be constructed analogously.
In further aspects of the invention, two, three or all four of the different aspects mentioned above are combined with each other. In a preferred combination, screening for proteases with decreased Michaelis-Menten constants is combined with the use of intermediate substrates. In another preferred combination, screening for proteases with decreased Michaelis-Menten constants is combined with the screening of two or more substrates consecutive or in parallel. In a further preferred combination, screening for proteases with decreased Michaelis-Menten constants is combined with using an excess of competing, unlabelled substrate. In a particularly preferred combination, the four aspects, screening for proteases with decreased Michaelis-Menten constants, screening of two or more substrates in parallel, the use of an excess of competing, unlabelled substrate, and the use of intermediate substrates, are combined with each other.
In a particularly preferred embodiment of the invention the target protease has a specificity similar to tissue-type plasminogen activator and cleaves the target substrate CPGR⇓VVGG. Such target protease can, among others, be generated by the above defined method of the invention when the starting protease is BARI protease from S. cerevisiae. In such method the following second/intermediate substrates are preferably utilized:
The present invention also pertains to the sequence-specific proteases obtainable by the methods described hereinbefore. In a preferred embodiment the sequence specific protease has a specificity similar to tissue-type plasminogen activator and cleaves the target substrate CPGR⇓VVGG. The starting protease preferably is BAR1 protease including, but not limited to, the one depicted in SEQ ID NO:8. Additionally, BAR1 proteases modified by truncation up to 200 aa, preferably in the range of 100 to 200 aa, more preferably in the range of 120 to 180 aa, most preferably in the range of 140 to 160 aa at the C- or N-terminal can be used as starting proteases. Even more preferably the sequence-specific protease is derived from said BAR1 derived protease and has at least one mutation selected from the group comprising the modifications L33I, Y45D, T47A, T59I, N82D, E96V, M107I, N123D, E143D, N151V, I152F, K161E, A163T, T165A, R178S, T221I, E231V, D321N, D367G, M369L, V370I, A399S, K404R and S440L. Particularly preferred among said proteases are those having at least one of D367G, V370I, M107I, I152F, E143D and E231V. The particularly preferred mutants of BAR1 protease may further be modified e.g. by truncation of up to 10 aa at the C- or N-terminal ends thereof or by deletion, insertion or substitution of up to 50 aa, preferably up to 20 aa, most preferably up to 10 aa within its sequence.
For the purpose of this figure, different shapes represent different amino acid residues, and the inverse profile of the shapes represent the protease's recognition sites, respectively. Shapes with a swung tilde at the top represent any amino acid residue at that position. The active site of the enzyme is indicated by an asterisk, and the arrow indicates the cleavage site within the substrate.
The type of the one or more proteases used as first proteases defines whether alternative A or alternative B is to be employed. In alternative A, the first protease is characterized by an already high specificity towards a defined, first substrate. According to the method of the invention, this specificity has to be changed qualitatively into the target specificity. In alternative B, the first protease has a relatively low specificity, i.e. it does not discriminate between a pool of substrates that differs, for example, at positions P2, P1′, and/or P2′. By the process of the invention, only the quantitative specificity of those proteases is increased towards the value of the target specificity.
These proteases are selected by screening with an excess of competing substrates (competitor approach) or by screening for higher substrate affinity (affinity approach). In general, the evolved protease can be identified by the comparison of the catalytic activity towards offered substrates including the first substrates and the target substrate.
The invention is further explained by the following Examples. It is understood that the examples and embodiments described therein are for illustrative purposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to included within the spirit and purview of this application and are considered within the scope of the appended claims. All publications, patents, and patent applications cited herein are hereby incorporated by reference in their entirety for all purposes.
In the following examples, materials and methods of the present invention are provided including the determination of catalytic properties of enzymes obtained by the method. It should be understood that these examples are for illustrative purpose only and are not to be construed as limiting this invention in any manner.
In the experimental examples described below, standard techniques of recombination DNA technology were used that were described in various publications, e.g. Sambrook et al (1989), Molecular Cloning; A Laboratory Manual, Cold Spring Harbor Laboratory, or Ausubel et al. (1987), Current Protocols in Molecular Biology 1987-1988, Wiley Interscience, Methods in Yeast Genetics (1994) A Cold Spring Harbour Laboratory Manual, which are incorporated herein in their entirety by reference. Unless otherwise indicated, restriction enzymes, polymerases and other enzymes as well as DNA purification kits were used according to the manufacturers specifications.
Genes encoding protease variants were cloned into a vector suitable for extracellular expression of proteins by the yeast Saccharomyces cerevisiae. The vector used is a derivate of the plasmid pYES2, which is commercially available from Invitrogen, Inc. A map of the plasmid is shown in
A population of protease variants was provided by random modification of genes encoding proteases with known substrate specificities, followed by expression of the protease variants encoded by these modified genes using S. cerevisiae as a suitable host organism. First, genes encoding protease variants with known substrate specificities were PCR amplified under error-prone conditions, essentially as described by Cadwell, R. C and Joyce, G. F. (PCR Methods Appl. 2 (1992) 28-33). Error-prone PCR was done using 30 pmol of each primer, 20 nmol dGTP and dATP, 100 nmol dCTP and dTTP, 20 fmol template, and 5 U Taq DNA polymerase in 10 mM Tris HCl pH 7.6, 50 mM KCl, 7 mM MgCl2, 0.5 mM MnCl2, 0.01% gelatin for 20 cycles of 1 min at 94° C., 1 min at 65° C. and 1 min at 72° C. The resulting DNA library was purified using the Qiaquick PCR Purification Kit following the suppliers' instructions, PCR products were digested with restriction enzymes XhoI and KpnI and purified as described in Example I. Afterwards, the PCR products were ligated into the vector which was digested with XhoI and KpnI, gel-purified and dephosphorylated. The ligation products were transformed into E. coli, amplified in LB containing ampicillin as marker, and the plasmids were purified using the Qiagen Plasmid Purification Kit following the suppliers' instructions. Resulting plasmids were transformed into S. cerevisiae cells. Populations of protease variants were provided by inducing expression in the transformed S. cerevisiae cells by adding 2% galactose to the medium.
Alternatively, genes encoding protease variants with known substrate specificities were statistically recombined at homologous positions by use of the Recombination Chain Reaction, essentially as described in WO 0134835. PCR products of the genes encoding the protease variants were purified using the QIAquick PCR Purification Kit following the suppliers' instructions, checked for correct size by agarose gel electrophoresis and mixed together in equimolar amounts. 80 μg of this PCR mix in 150 mM TrisHCL pH 7.6, 6.6 mM MgCl2 were heated for 5 min at 94° C. and subsequently cooled down to 37° C. at 0.05° C./sec in order to re-anneal strands and thereby produce heteroduplices in a stochastic manner. Then, 2.5 U Exonuclease III per μg DNA were added and incubated for 20, 40 or 60 min at 37° C. in order to digest different lengths from both 3′ ends of the heteroduplices. The partly digested PCR products were refilled with 0.6 U Pfu polymerase per μg DNA by incubating for 15 min at 72° C. in 0.17 mM dNTPs and Pfu polymerase buffer according to the suppliers' instructions. After performing a single PCR cycle, the resulting DNA was purified using the QIAquick PCR Purification Kit following the suppliers' instructions, digested with KpnI and XhoI and ligated into the linearized vector. The ligation products were transformed into E. coli, amplified in LB containing ampicillin as marker, and the plasmids were purified using the Qiagen Plasmid Purification Kit following the suppliers' instructions. Resulting plasmids were transformed into S. cerevisiae cells. Populations of protease variants were provided by inducing expression in the transformed S. cerevisiae cells by adding 2% galactose to the medium.
All peptide substrates were synthesized on a peptide synthesizer using the approach of Merrifield et al. (Nature. 207 (1965) 522-523). Peptide substrates that resemble the target substrate were designed by substituting the amino acid residues at one or more positions of the first peptide substrate with the amino acid residues at the one or more positions of the target substrate. Alternatively, the amino acid residues at one or more positions of the first peptide substrate were substituted with amino acid residues that have an intermediate character with respect to the amino acid residues of the first peptide substrate and the amino acid residues of the target peptide substrate. For the determination of the intermediate character of amino acid residues refer to Table III, Marker fluorophores were attached to the peptide substrates either via the amino group of the N-terminus or via the carboxy group of the C-terminus. Alternatively, a cysteine residue was added either at the N-terminus or at the C-terminus of the peptide, and the marker fluorophor was chemically attached to the thiol group of the cysteine residue.
Alexa 488 (Molecular Probes Inc., Oregon, USA) and Cy-5 (Amersham Biosciences Europe GmbH, Freiburg, Germany) were typically used as fluorophor markers. Protease cleavage of the peptide substrate was monitored by cross-correlation FCS (Proc. Natl. Acad. $ci. USA. 95 (1998) 1416-1420). As an example, the cleavage of a peptide substrate that contains the target substrate for tobacco etch virus protease (TEV protease) and has the Alexa 488 fluorophor attached to the C-terminus of the peptide and the Cy-5 fluorophor attached to the N-terminus of the peptide is shown in
In order to identify enzyme variants having the desired substrate specificity, a screening approach based on a confocal fluorescence spectroscopy set-up as disclosed in WO 9416313 was used. Either the cell suspension of a S. cerevisiae culture directly, or an aliquot of the cell-free supernatant was used as the sample containing the secreted protease variant. After adding the substrate to the sample and incubation for a certain period of time, the samples were subjected to measurement by confocal fluorescence spectroscopy. If necessary, this procedure was repeated several times in order to measure kinetics of the proteolytic cleavage. Consequently, the samples were ranked according to proteolytic activity, and samples exceeding a certain activity threshold were identified in order to isolate the gene encoding the corresponding protease variant. The distribution of proteolytic activities of protease variants obtained by this procedure is shown in
Protease variants that have an increased affinity towards the target peptide substrate were generated by the method of the invention based on screening at low substrate concentrations. By means of error-prone PCR (according to Cadwell, R. C. and Joyce, G. F., PCR Methods Appl. 2 (1992) 28-33), a population of protease variants was generated that is related to the alkaline protease subtilisin E from Bacillus subtilis, which has a relatively low specificity (s=0.82). This correlates to the relatively high KM which is in the range of 150-200 μM. The population of protease variants was screened at a complexity of 106 variants by confocal fluorescence spectroscopy employing substrate concentrations in the range of 10 nM. Variants isolated in this first screen were used as first proteases in a second cycle to provide another population of protease variants. Analogously, variants isolated in subsequent cycles were used as first proteases in the following cycle. The population of variants provided in the second cycle and all subsequent cycles was generated by a combination of error-prone PCR (see above) and in-vitro homologous recombination (according to WO 0134835). Variants isolated from the first four cycles of this procedure were analyzed kinetically. The increase in affinity towards the substrate over the four rounds corresponds to the decrease in KM of the best performers of each cycle which is shown in
Proteases were generated by the method of the invention that had a specificity that was altered towards the specificity of tissue-type plasminogen activator (t-PA). The BAR1 protease from Saccharomyces cerevisiae (SEQ ID NO:8) was used as first protease. This protease belongs to the group of aspartic proteinases (Mac Kay et al.; Structure an Function of the Aspartic Proteinases (1991) 161-172). It is specific for peptide substrates containing the amino acid sequence WLQLKPGQ, and catalyses the cleavage at the peptide bond between the second leucine and the lysine residue. Populations of protease variants that were related to the BAR 1 protease or proteases isolated in subsequent screening cycles were generated by means of error-prone PCR (according to Cadwell, R. C and Joyce, G. F., PCR Methods Appl. 2 (1992) 28-33) and in-vitro homologous recombination using the Recombination Chain Reaction (WO 0134835). Protease variants were screened for proteolytic activity at complexities of 106 variants by confocal fluorescence spectroscopy. The BAR 1 protease used as first protease already had a relatively high specificity which was in the range of the target specificity. Therefore, a combination of the affinity approach and of the intermediate approach was used. Screening at low concentrations kept the specificity of the protease high, while screening on intermediate substrates enabled the evolution towards the new specificity. Four intermediate substrates were constructed. Intermediate substrate 1 had the amino acid sequence WLGLVPGG, intermediate substrate 2 the amino acid sequence WLGQVPGG, intermediate substrate 3 the amino acid sequence WLGRVPGG, and intermediate four had the sequence WLGRVVGG. The target substrate specificity of t-PA is directed to CPGRVVGG with cleavage between the arginine residue and the first valine residue. All substrates are shown in Table IV.
Intermediate 1 was an amino acid composition intermediate due to the fact that it contained at positions P4, P3, P1 and P2′ the same amino acid residues as the first substrate, and at positions P2, P1′, and P4′ the same residues as the target substrate. Intermediate 2 was an amino acid property intermediate with regard to intermediate 1 and the target substrate. It resembled intermediate 1 but contained at position P1 a glutamine residue which has an intermediate character compared to the leucine residue present at that position in the first substrate and the arginine residue present at that position in the target substrate. Intermediate 3 as another amino acid composition intermediate was based on amino acid residues stemming from both, the first substrate and the target substrate, as intermediate substrate 1 does, but, in contrast to the latter one, shared one additional position with the target substrate. Compared to intermediate 3, intermediate 4 shares one further amino acid with the target sequence at position P2′. The changed specificities of different variants that were generated by this method are shown in
Increase of substrate specificity can also be measured as time-dependent conversion of the substrates, as exemplarily demonstrated in
In this way proteases are generated according to the method of the invention, that have a substrate specificity similar to the human tissue-type plasminogen activator. The proteases generated have at least one mutation at a position out of the group; 33, 45, 47, 59, 82, 96, 107, 123, 143, 151, 152, 161, 163, 165, 178, 221, 231, 321, 367, 369, 370, 399, 404, 440 (based on the numbering of the amino acid sequence of the protease BAR1 listed as SEQ ID NO:8) Preferably, a protease variant evolved from the BAR1 wt protease towards specificity of the human tissue-type plasminogen activator has at least one mutation out of the group: D367G, M369L, V370I, M107I, I152F, E143D, E231V, L33I, Y45D, T47A, T59I, N82D, E96V, N123D, N151V, K161E, A163T, T165A, R178S, T221I, D321N, A399S, K404R and S440L.
| Number | Date | Country | |
|---|---|---|---|
| 60379504 | May 2002 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 10435754 | May 2003 | US |
| Child | 11746145 | US |
