AU2004231193B2 - Chemically modified enzymes with multiple charged variants - Google Patents

Description

AUSTRALIA

Patents Act 1990 GENENCOR INTERNATIONAL, INC.

COMPLETE SPECIFICATION STANDARD PATENT Invention Title.

Chemically modified enzymes with multiple charged variants The following statement is a full description of this invention including the best method of performing it known to us:r Sla N CHEMICALLY MODIFIED ENZYMES WITH MULTIPLE CHARGED O

VARIANTS

oo This application is a divisional application of Australian Patent Application No.

22001/00, of which the entire contents are incorporated herein by reference.

FIELD OF THE INVENTION This invention relates to the field of chemically modified mutant enzymes. In particular this invention pertains to chemically modified mutant enzymes in which multiply O 10 charged substituents are introduced to enhance interaction of the enzyme with a ,I charged substrate.

BACKGROUND OF THE INVENTION For both protein chemistry (Nilsson et al. (1992) Curr. Opin. Struct. Biol 2: 569-575; LaVallie and McCoy (1995) Curr. Opin. Biotechnol. 6: 501-506; Uhlen and Moks (1990) Methods Enzymol. 185: 129-143) and organic synthesis applications (Sears and Wong (1996) Biotechnol. Prog. 12: 423-433; Faber (1997) Biotransformations in Organic Synthesis: 3rd ed. Springer-Verlag: Heidelberg; Roberts (1993) Preparative Biotransformations; Wiley: New York: 1993) it is desirable to have available a diverse toolbox of inexpensive proteases with high selectivity and diverse substrate preferences. To date the most extensively exploited class of enzymes in organic synthesis applications have been the hydrolases. Among these, the serine proteases have received considerable attention due, in part, to their often exquisite stereo-, regio-, and chemo-selectivities (Sears and Wong (1996) Biotechnol. Prog. 12: 423-433; Faber (1997) Biotransformations in Organic Synthesis: 3rd ed. Springer-Verlag: Heidelberg; Roberts (1993) Preparative m:\specifications\090000\93000\9384Ocmphxgdiv.doc WO-0037658 P)PCT/US99/30362 Biotransformations; Wiley: New York: 1993; Moree et al. (1997) J. Am. Chem. Soc. 119: 3942-3947).

While over 3000 enzymes have now been reported, of which many are 0 proteases, significantly fewer of the latter are available inexpensively from commercial 0 5 sources (Faber (1997) Biotransformations in Organic Synthesis: 3rd ed. Springer-Verlag: Heidelberg; Roberts (1993) Preparative Biotransformations; Wiley: New York: 1993; Moree et al. (1997) J. Am. Chem. Soc. 119: 3942-3947; Jones (1986) Tetrahedron 42: 3351- S 3403). Furthermore, since wild type enzymes do not accept all substrate structures of synthetic interest, it is attractive to contemplate the tailoring of a readily available protease in order to expand their substrate specificities in a controlled manner with the ultimate goal of C creating any desired specificity at will.

In this regard, the goal of specificity alteration of enzymes has already been targeted by several different approaches. For example, site-directed mutagenesis (Perona and Craik (1995) Protein Sci. 4: 337-360) and random mutagenesis (Arnold (1998) Acc. i Chem. Res. 31(3): 125-131) have been employed to tailor enzyme specificity and have permitted some insights into the electrostatic (Wells et al. (1987) Proc. Natl. Acad. Sci. USA, 84: 5167-5171; Wells et al. (1987) Proc. Nat. Acad. Sci. USA, 84: 1219-1223; Wells and Estell (1988) TIBS 13: 291-297; Bott et al. (1987) Pages 139-147 In: Biotech. Agric.Chem.; Lebanon, Mumma, Honeycutt, Duesing, eds.; Vol. ACS Symp. Ser. 334; Russell et al.

(1987) J. Mol. Biol. 193: 803-813; Ballinger et al. (196) Biochemistry 33: 13579-13585), steric (Rheinnecker et al. (1994) Biochemistry 33:221-225; Rheinnecker et al. (1993) Biochemistry 32(5): 1199-1203; Sorensen et al. (1993) Biochemistry 32: 8994-8999; Estell et al. (1986) Science 233: 659-663; Takagi et al. (1996) FEBSLett. 395: 127-132; Takagi et al.

I (1997) Protein Eng. 10(9): 985-989), and hydrophobic (Estell et al. (1986) Science 233: 659- 663; Wangikar et al. (1995) Biochemistry 34(38): 12302-12310; Bech et al. (1993) Biochemistry 32: 2845-2852) factors which govern enzyme-substrate interactions. However, the structural variations within these approaches are limited to the 20-natural amino acids.

Consequently, biosynthetic methods have recently been developed to introduce unnatural amino acids into proteins (25. Comish et al. (1995) Angew. Chem. Int. Ed.Eng. 34: 621-633; Parsons et al. (1998) Biochemistry 37: 6286-6294; Hohsaka et al. (1996) J. Am. Chem. Soc.

118(40): 9778-9779). Unnatural functionalities have also been incorporated by chemical modification techniques (Kuang et al. (1996) J. Am. Chem. Soc. 118: 10702-10706; Ory et al. (1998) Protein. Eng. 11(4): 253-261; Peterson: E. Hilvert: D. Biochemistry 34: 6616- -2-

I

00 WO00/37658 PCT/US99/30362 6620; Suckling: C. Zhu: Bioorg. Med. Chem; Lett. 3: 531-534; Rokita and Kaiser (1986) J. Am. Chem. Soc. 108: 4984-4987; Kokubo et al. (1987) J. Am. Chem. Soc. 109: 606-607; Radziejewski et al. (1985) J. Am. Chem. Soc. 107: 3352-3354). Generally, however, unnatural amino acid mutagenesis approach is not yet amenable to large scale preparations, and chemical modification alone is insufficiently specific.

SUMMARY OF THE INVENTION This invention provides novel multiply-charged chemically modified mutant J ,enzymes. In a particularly preferred embodiment this invention provides a modified enzyme where one or more amino acid residues in the enzyme are replaced by cysteine residues. The 10 cysteine residues are modified by replacing the thiol hydrogen in the residue with a substituent group providing a thiol side chain comprising a multiply charged moiety.

Preferred enzymes include serine hydrolases, more preferably proteases subtilisins).

One particularly preferred enzyme is a Bacillus lentus subtilisin. The amino acid replaced with a cysteine may include an amino acid selected from the group consisting of asparagine, leucine, methionine, and serine. Preferred replaced amino acids are in a binding site a subsite such as S1, Sl', and S2). Where the enzyme is a subtilisin-type serine hydrojase the cysteine(s) is substituted amino acid(s) corresponding to a Bacillus lentus subtilisin residue selected from the group consisting of residue 156, reside 166, residue 217, residue 222, residue 62, residue 96, residue 104, residue 107, reside 189, and residue 209. Where the enzyme is a trypsin-chymotrypsin-type serine protease the cysteine(s) are substituted for and amino acid corresponding to a trypsin residue selected from the group consisting ofTyr94, Leu99, Gln175, Aspl89, Serl90, and Gln192. Where the enzyme is an alpha/beta serine hydrolase the cysteine(s) are substituted for and amino acid corresponding to a Candida antartica lipase (protein Data Bank entry Itca) residue selected from the group consisting of Trpl04, Thrl38, Leul44, Vall54, Ile189, Ala 225, Leu278 and Ile185.

The multiply charged moiety can be negatively or positive charged and in certain embodiments, the enzyme can contain both positively and negatively multiply charged moieties. Particularly preferred negatively charged moieties include, but are not limited to, sulfonatoethyl thiol, 4-carboxybutyl thiol, 3,5-dicarboxybenzyl thiol, 3,3dicarboxybutyl thiol, and 3,3,4-tricarboxybutyl thiol, while particularly preferred positively charged moieties include, but are not limited to, aminoethyl thiol, 2- -3- SWOb8/37658 PCT/US99/303 6 2 (trimethylammonium)ethyl thiol, 4,4-bis(aminomethyl)-3-oxo-hexyl thiol, and 2,2- Sbis(aminomethyl)-3-aminopropyl thiol. The multiply charged moiety can also be a Sdendrimer or a polymer.

0 In another embodiment, this invention provides methods of making novel multiply-charged chemically modified mutant enzymes. The methods involve providing an 00 enzyme having one or more amino acids have been replaced with cysteine residues; and replacing the thiol hydrogen, in one or more cysteine residues, with a substituent group Sproviding a thiol side chain comprising a multiply charged moiety. In certain embodiments, Sa native cysteine can be chemically modified and there is no need to introduce a cysteine.

10 Preferred enzymes include serine hydrolases as identified herein. Preferred residues for replacement with a cysteine and preferred multiply-charged moieties are identified herein.

In another embodiment, this invention includes a composition comprising any Sone of the multiply charged chemically modified mutant enzymes as described herein and a detergent or other cleaning agent.

In still another embodiment, this invention provides methods of assaying for a preferred enzyme. The methods involve providing a swatch of material comprising a piece of material and a stain; fixing the stain to the material; applying an enzyme to the swatch; and incubating the watch and the enzyme. The method can further involve determining the degree of removal of the stain from the material. Preferred enzymes for use in this method include, but are not limited to proteases, a cellulases, amylases, laccases, and lipases.. In particularly preferred embodiments, the enzymes are modified serine hydrolases as described herein. Preferred materials include, but are not limited to fabrics, plastics, or ceramics.

Preferred stains include, but are not limited to blood, milk, ink, grass, gravy, chocolate, egg, cheese, clay, pigment, and oil. One particularly preferred stain is a blood/milk/ink (BMI) stain.

The method can also involve incubating the stain with a cross-linking agent hydrogen peroxide, bleaching agents, glutaraldehyde, and carbodiimides). The enzyme can be applied to the swatch in combination with a detergent ingredient. The method can additionally involve agitating the swatch and enzyme during incubation.

In still yet another embodiment, this invention provides methods of assaying for a preferred detergent composition. These methods involve providing a swatch of material comprising a piece of material and a stain; fixing the stain to the material; applying a detergent composition to the swatch; and incubating the watch and the detergent -4- 00 composition. The methods can additionally involve determining the degree of removal t of the stain from the material. Preferred enzymes, materials and stains are as described Sherein. The method can involve incubating the stain with across-linking agent(e. g., hydrogen peroxide, bleaching agents, glutaraldehyde, and carbodiimides). The enzyme can be applied to the swatch in combination with the enzyme. In certain embodiments, Cc the method involves agitating the swatch and detergent composition during incubation.

_This invention also provides methods of determining the catalytic efficiency of an enzyme. The methods involve providing a swatch of material comprising a piece of material and a stain; applying the enzyme to the swatch; incubating the swatch and the enzyme; removing the swatch or supernatant; and measuring a constituent of the stain.

CI Preferred enzymes, materials and stains are as described herein. The method can involve incubating the stain with a cross-linking agent hydrogen peroxide, bleaching agents, glutaraldehyde, and carbodiimides). In certain preferred embodiments, the constituent is in the supernatant. The constituent can be measured by determining its fluorescenceand/or absorbance g. absorbance spectra).

Also included herein are kits for the practice of the methods of this invention.

One kit comprises a container containing a modified enzyme where one or more amino acid residues in the enzyme are replaced by cysteine residues, and the cysteine residues are modified by replacing the thiol hydrogen in the cysteine residues with a substituent group providing a thiol side chain comprising a multiply charged moiety. Another kit comprises a container containing a methane sulfonate reagent comprising a multiply charged substituent, and instructional materials teaching the use of the sulfonate reagent to couple a mutiplycharged moiety to a cysteine residue in a protein.

The present invention provides a modified enzyme wherein one or more amino acid residues in the enzyme are replaced by cysteine residues, wherein the cysteine residues are modified by replacing the thiol hydrogen in said cysteine residues with a substituent group providing a thiol side chain comprising a multiply charged moiety.

The present invention also provides a method of producing a modified enzyme, said method comprising: providing an enzyme wherein one or more amino acids have been replaced with cysteine residues; and replacing the thiol hydrogen, in said one or more cysteine residues, with a substituent group providing a thiol side chain comprising a multiply charged moiety.

Any discussion of documents, acts, materials, devices, articles or the like which has been included in the present specification is solely for the purpose of providing a context for the present invention. It is not to be taken as an admission that any or all of 00 these matters form part of the prior art base or were common general knowledge in the tfield relevant to the present invention as it existed before the priority date of each claim Sof this application.

Throughout this specification the word "comprise", or variations such as "comprises" or "comprising", will be understood to imply the inclusion of a stated Cc element, integer or step, or group of elements, integers or steps, but not the exclusion of any other element, integer or step, or group of elements, integers or steps.

N DEFINITIONS The terms "polypeptide", "peptide" and "protein" are used interchangeably herein to refer to a polymer of amino acid residues. The terms apply to amino acid polymers in which one or more amino acid residue is an artificial chemical analogue of a corresponding naturally occurring amino acid, as well as to naturally occurring amino acid polymers. The term may also include variants on the traditional peptide linkage joining the amino acids making up the polypeptide.

The term"residue"as used herein refers to natural, synthetic, or modified amino acids.

I PCT/US99/30362 WO 00/37658 The terms enzyme includes proteins that are capable of catalyzing chemical changes in other substances without being permanently changed themselves. The enzymes Scan be wild-type enzymes or variant enzymes. Enzymes within the scope of the present Z invention include, but are not limited to, pullulanases, proteases, cellulases, amylases, 00 5 isomerases, lipases, oxidases, oxidoreductases, hydrolases, aldolases, ketolases, Sglycosidases, oxidoreductases, hydrolases, aldolases, ketolases, glycosidases, lyases, ligases, transferases, and ligases.

m The phrase "multiply-charged" or "multiple-charge" refers to a net charge greater than +1 or less than -1 at pH 7.0. A multiply charged substituent is a substituent that when covalently coupled to a subject enzyme bears a net charge greater than +1 or less than 1 at pH A "mutant enzyme" is an enzyme that has been changed by replacing an Samino acid residue with a cysteine (or other) residue.

A "chemically modified" enzyme is an enzyme that has been derivatized to bear a substituent not normally found at that location in the enzyme.

A "chemically modified mutant enzyme" or "CMM" is an enzyme in which an amino acid residue has been replaced with another amino acid residue (preferably a cysteine) and the replacement residue is chemically derivatized to bear a substituent not normally found on that residue.

The term "thiol side chain group", "thiol containing group", and thiol side chain" are terms that can be used interchangeably and include groups that are used to replace the thiol hydrogen of a cysteine. Commonly the thiol side chain group includes a sulfur atom through which the thiol side chain group is attached to the thiol sulfur of the cysteine.

The "substitutent" typically refers to the group remains attached to the cysteine through a disulfide linkage formed by reacting the cysteine with a methanesulfonate reagent as described herein. While the term substituent preferably refers just to the group that remains attached (excluding its thiol group), the substituent can also refer to the entire thiol side chain group. The difference will be clear from the context.

The "binding site of an enzyme" consists of a series of subsites across the surface of the enzyme. The substrate residues that correspond to the subsites are labeled

P

and the subsites are labeled S. By convention, the subsites are labeled S 1 S2, S3, S 4 SI', and

S

2 A discussion of subsites can be found in Siezen et al. (1991) Protein Engineering, 4: 0 SW00/37658 PCT/US99/303 6 2 719-737, and Fersht (1985) Enzyme Structure and Mechanism, 2nd ed. Freeman, New York, C29-30. The preferred subsites include Si, S1', and S 2

C

The terms "stereoselectivity" or "stereoselective" when used in reference to an O enzyme or to a reaction catalyzed by an enzyme refers to a bias in the amount or C' 5 concentration of reaction products in favor of enantiomers of one chirality. Thus a stereoselective reaction or enzyme will produce reaction products that predominate in the form over the form (or form over the form) or conversely that predominate in the form over the form (or form over the form). The predominance of one chirality is preferably a detectable predominance, more preferably a substantial I 10 predominance, and most preferably a statistically significant predominance at a confidence level of at least 80%, preferably at least 90%, more preferably at least 95%, and most preferably at least 98%).

The phrase amino acid or "amino acid in the XX subsite" is intended to include the amino acid at the referenced position amino 156 of B. lentus subtilisin which is in the Si subsite) and the amino acids at the corresponding (homologous) position in related enzymes.

A residue (amino acid) of a enzyme is equivalent to a residue of a referenced enzyme B. amyloliquefaciens subtilisin) if it is either homologous corresponding in position in either primary or tertiary structure) or analagous to a specific residue or portion of that residue in B. amyloliquefaciens subtilisin having the same or similar functional capacity to combine, react, or interact chemically).

In order to establish homology to primary structure, the amino acid sequence of the subject enzyme a serine hydrolase) is directly compared to a reference enzyme S(e.g. B. amyloliquefaciens subtilisin) primary sequence and particularly to a set of residues known to be invariant in all enzymes of that family (e.g subtilisins) for which sequence is Sknown. After aligning the conserved residues, allowing for necessary insertions and deletions in order to maintain alignment avoiding the elimination of conserved residues through arbitrary deletion and insertion), the residues equivalent to particular amino acids in the primary sequence of the reference enzyme B. amyloliquefaciens subtilisin) are defined. Alignment of conserved residues preferably should conserve 100% of such residues. However, alignment of greater than 75% or as little as 50% of conserved residues is also adequate to define equivalent residues. Conservation of the catalytic triad, Asp32/His64/Ser221) should be maintained. WO'00/37658 PCTUS99/30362 SThe conserved residues may be used to define the corresponding equivalent Oamino acid residues in other related enzymes. For example, the two (reference and "target") sequences are aligned in to produce the maximum homology of conserved residues. There 0 may be a number of insertions and deletions in the "target" sequence as compared to the 00 5 reference sequence. Thus, for example, a number of deletions are seen in the thermitase sequence as compared to B. amyloliquefaciens subtilisin (see, e.g. U.S. Patent 5,972,682).

Thus, the equivalent amino acid or Tyr217 in B. amyloliquefaciens subtilisin in thermitase is C the particular lysine shown beneath Tyr217.

The particular "equivalent" resides may be substituted by a different amino T 10 acid to produce a mutant carbonyl hydrolase since they are equivalent in primary structure.

CEquivalent residues homologous at the level of tertiary structure for a 0particular enzyme whose tertiary structure has been determined by x-ray crystallography, are C defined as those for which the atomic coordinates of 2 or more of the main chain atoms of a particular amino acid residue of the reference sequence B. amyloliquefaciens subtilisin) i 15 and the sequence in question (target sequence) (N on N, CA on CA, C on C, and O on O) are within 0.13 nm and preferably 0.1 nm after alignment. Alignment is achieved after the best model has been oriented and positioned to give the maximum overlap of atomic coordinates of non-hydrogen protein atoms of the enzyme in question to the reference sequence. The best model is the crystallographic model giving the lowest R factor for experimental diffraction data at the highest resolution available.

7-fo(h)l Ifc(h)l h

R=

E Ifo(h)l h Equivalent residues which are functionally analogous to a specific residue of a reference sequence B. amyloliquefaciens subtilisin) are defined as those amino acids sequence in question related subtilisin) which may adopt a conformation such that they will alter, modify or contribute to protein structure, substrate binding or catalysis in a manner defined and attributed to a specific residue of the reference sequence as described herein. Further, they are those residues of the sequence in question (for which a tertiary structure has been obtained by x-ray crystallography), which occupy an analogous position to the extent that although the main chain atoms of the given residue may not satisfy the criteria of 00 0 WO'00/37658 PCT/US99/30362 equivalence on the basis of occupying a homologous position, the atomic coordinates of at least two of the side chain atoms of the residue lie with 0.13 nm of the corresponding side C(7 chain atoms of the reference sequence. The three dimensional structures would be aligned as outlined above. For an illustration of this procedure see U.S. Patent 5,972,682.

i 5 A "serine hydrolase" is a hydrolytic enzyme utilizing an active serine side chain to serve as a nucleophile in a hydrolytic reaction. This term includes native and synthetic serine hydrolases as well as enzymes engineered to perform the reverse reaction, for synthetic purposes.

;The "alpha/beta serine hydrolases" are a family of serine hydrolyases based on structural homology to enzymes including wheat germ serine carboxypeptidase's II (see, Liam et al. (1992) Biochemistry 31: 9796-9812; Olli's et al. (1992) Protein Engineering, 197-211).

The "subtilisin type serine proteases" refer to a family of serine hydrolyases based on structural homology to enzymes in including subtilisin BPN' (Bott et al. (1988) J.

Biol. Chem. 263: 7895-7906; Siezen and Louise (1997) Protein Science 6: 501-523).

Subtilisin are bacterial or fungal proteases which generally act to cleave peptide bonds of proteins or peptides. As used herein, "subtilisin" means a naturally-occurring subtilisin or a recombinant subtilisin. A series of naturally-occurring subtilisins is known to be produced and often secreted by various microbial species. Amino acid sequences of the members of this series are not entirely homologous. However, the subtilisins in this series exhibit the same or similar type ofproteolytic activity. This class of serine proteases shares a common amino acid sequence defining a catalytic triad which distinguishes them from the chymotrypsin related class of serine proteases. The subtilisins and chymotrypsin related serine proteases have a catalytic triad comprising aspartate, histidine and serine. In the S 25 subtilisin related proteases the relative order of these amino acids, reading from the amino to carboxy terminus, is aspartate-histidine-serine. In the chymotrypsin related proteases, the relative order, however, is histidine-aspartate-serine. Thus, subtilisin herein refers to a serine protease having the catalytic triad of subtilisin related proteases.

The "chymotrypsin serine protease family" refers to a family of serine hydrolyases based on structural homology to enzymes including gamma chymotrypsin (Birktoft and Blow (1972) J. Molecular Biology 68: 187-240).

A "dendritic polymer" is a polymer exhibiting regular dendritic branching, formed by the sequential or generational addition of branched layers to or from a core. The -9i WO 00/37658 PCT/US99/30362 term dendritic polymer encompasses "dendrimers", which are characterized by a core, at Sleast one interior branched layer, and a surface branched layer (see, Petar et al. Pages

C

641-645 In Chem. in Britain, (August 1994). A "dendron" is a species of dendrimer having 0 branches emanating from a focal point which is or can be joined to a core, either directly or Z 5 through a linking moiety to form a dendrimer. Many dendrimers comprise two or more 00 dendrons joined to a common core. However, the term dendrimer is used broadly to encompass a single dendron.

Dendritic polymers include, but are not limited to, symmetrical and unsymmetrical branching dendrimers, cascade molecules, arborols, and the like, though the most preferred dendritic polymers are dense star polymers. The PAMAM dense star C dendrimers (disclosed in U.S. Patent 5,714,166) are symmetric, in that the branch arms are of equal length. The branching occurs at the hydrogen atoms of a terminal -NH 2 group on a 0preceding generation branch. The lysine-based dendrimers are unsymmetric, in that the branch arms are of a different length. One branch occurs at the epsilon nitrogen of the lysine molecule, while another branch occurs at the alpha nitrogen, adjacent to the reactive carboxy group which attaches the branch to a previous generation branch.

Even though not formed by regular sequential addition of branched layers, hyperbranched polymers, hyperbranched polyols, may be equivalent to a dendritic polymer where the branching pattern exhibits a degree of regularity approaching that of a dendrimer.

A "swatch" is a piece of material a natural or synthetic textile) that has a stain applied thereto. The material can be, for example, a fabric made of a natural fiber (e.g.

cotton, hemp, wool), or a synthetic material nylon, polyester, rayon, etc.) or a mixture of synthetic or natural fibers. The material need not be a textile, but can be any material subject to cleaning operations glass, ceramic, Formica, etc.).

SA "smaller swatch" is one that has been cut from the swatch of material either before or after fixing a stain to he swatch and can, for example, fit into the well ofa 48 or 96 well microtiter plate. The "smaller swatch" can also be made by applying a stain to a smaller piece of material. Preferably the smaller swatch is about 5/8 inch in diameter, more preferably the smaller swatch is about 0.25 inches in diameter.

The phrase" replacing the thiol hydrogen, in said one or more cysteine residues" does not require that every thiol hydrogen in every cysteine residue be replaced.

0

O

0

O

WO 09/37658 PCT/US99/30362 BRIEF DESCRIPTION OF THE DRAWINGS SFigure 1 illustrates the active site of SBL and shows the residues of the C, catalytic triad, Ser221, His64 and Asp32. The irreversible phenylmethylsulfonyl inhibitor 0 (pink) forms a bond to the O atom of Ser221 and its phenyl ring occupies the St binding site. The residues chosen for mutation, Asn62 in the S2 site, Ser156 and Ser166 in the S1 site and Leu217 in the SI' site are highlighted.

Figure 2 shows scheme 2 for the preparation of the aromatic dicarboxylate MTS Ic from toluene-3,5-dicarboxylic acid via a precursor benzylic bromide 3. (i) i Im 2 CO, DMF, 400 C then DBU, t-BuOH, 84% (ii) NBS, azobis(cyclohexanecarbonitrile), CC14, A, 96% NaSSO 2

CH

3 DMF, 50° C, 60% CF 3 COOH, CH 2 C2, 91%.

Figure 3 shows scheme 3 illustrating the preparation of the aliphatic di- and tricarboiylates ld,e were prepared from Meldrum's acid (5a) using 1,2-dibromoethane to introduce a brominated linker group. K 2

CO

3 DMF then BrCH 2 COOBu', 59% (b) K2C03, DMF then Br(CH 2 2 Br, 71% for 6b, 66% for 6c NaSSO 2

CH

3 DMF, 50° C, 83% for 7b, 86% for 7c Dowex 50W(H), p-dioxan, H 2 0, 79% CF 3 COOD, D 2 0, 500 C, Figures 4A, 4B, 4C, and 4D illustrate altered specificity patterns relative to WT as the level of negative charge increases in N62C, L271C, S156C and S166C mutants and CMMs with suc-AAPF-pNA as the substrate: Figure 4A: The kct/KMs for N62C CMMs alternate at moderately reduced levels, 1.5- to 3.5-fold lower than WT, which are established by the initial mutation to N62C Figure 4B: L217C CMMs show steady but lower levels ofkca/Ku, 4- to 5.5-fold lower than WT, which are again established by the initial mutation to cysteine. The exception is L217C-c which is only 2.5-fold lower than WT, possibly due to favorable binding of substrate to the phenyl ring of the aromatic side chain introduced by modification. Figure 4C: From the small reduction caused by mutation to S 156C kc/Kus decrease monotonically to 6-fold lower than WT for S156C-d. The kca/KM of S156C-e is partially restored. Figure 4D: kcaI/Ku decreases only 2.5-fold upon mutation to S 166C but decreases dramatically to 11-fold lower than WT when the negatively charged sulfonatoethyl side chain a is introduced. In parallel to N62C and L217C CMMs, k,./Ku for S166C CMMs does not decrease further to any significant extent as the level of negative charge increases.

W 00/37658' PCT/US99/30362 Figure 5 illustrates the effects of introducing negative charge to CMMs: In S(kca,/Ku), with suc-AAPF-pNA as the substrate, of the negatively charged N62C, L217C and c S 156C CMMs decreases relative to that of near-isosteric uncharged CMMs as the level of 0 negative charge increases (from side chain a to In contrast, this value for-the Z 5 corresponding S166C CMMs increases with increasing negative charge.

00 Figure 6 shows scheme 3 for the synthesis of the doubly-charged MTS reagent Ic from the readily available dichloroalcohol 1 NaN 3 1300 C, 98% NaH, SBu 4 NI, THF then 4, 55% H 2 Pd-black, MeOH Boc20, NaOH(aq.)/dioxan, 72% over 2 steps TBAF, THF, 93% MsCI, Et 3 N, CH2C12, 99% LiBr, acetone, A, 93% (h) NaSSO 2 CH3, DMF, 500 C, 87% CF 3 COOH, CH 2 C1 2 then ion exchange chromatography, C'-i 67%.

Figure 7 show scheme 4 for the synthesis of triply-charged MTS Id from ci pentaerythritol HBr, AcOH, A then HBr, c. H 2

SO

4 A, 49% NaN 3 DMF, 100* C (c) PPh 3

NH

3 (aq)/dioxan (ii) HC1 57% over 3 steps PhCHO, Et 3 N, MeOH, 99% (e) MsCI, Et 3 N, CH2C12, 77% LiBr, acetone, A, 78% NaSSO 2

CH

3 DMF, 800 C, 65% (d) HCI 69%.

Figures 8A, 8B, 8C and 8D show altered specificity patterns for N62C (Fig.

8A), L271C (Fig. 8B), S156C (Fig. 8C), and S166C (Fig. 8D) CMMs: variations in In (kca/Ku), with suc-AAPF-pNA as the substrate, for cysteine mutants and positively charged CMMs relative to WT.

Figure 9 show scheme 5 for the preparation of chemically modified mutant enzymes (CMMs) having multiply charged substituents. This approach entails the introduction of a unique cysteine residue at a selected position, followed by its chemical modification with methanethiosulfonate reagents MTS, la-1j) to generate chemically modified mutant enzymes (CMMs).

Figure 10 shows the klc/K screen of WT-SBL and S166C-S-a to -c CMMs with each of the standard suc-AAPF-pNA substrate and with the suc-AAP-A/R/E-pNA substrates.

Figure 11 shows a plot of side-chain negative charge on R at position 166 versus kt/KM with suc-AAPR-pNA. The points on the line are for the aliphatic series of carboxylated -R groups and the WT, Ser166 WO 00/37658 PCT/US99/30362 DETAILED DESCRIPTION This invention provides chemically modified mutant enzymes (CMMs) that bear multiply charged moieties. Without being bound to a particular theory, it was discovered that the addition of charge (positive or negative) improves the interaction of an enzyme with (oppositely) charged substrates. Moreover, it was a discovery of this invention that increasing the charge on the enzyme can increase the specificity and/or activity of that enzyme for a substrate, particularly for a charged substrate.

This is illustrated herein in Example 3 wherein a series of mono-, di-, and trinegatively charged chemically modified mutant enzymes generated and all showed improved kca/KMs with the positively charged PI residue containing substrate, suc-AAPRpNA. Moreover, virtually arithmetic improvements in kcat/K were exhibited with increasing number of negative charges on the S 166C-R side chain. These increases culminated in a 9-fold improvement in k.atKM for the suc-AAPR-pNA substrate and a 61fold improvement in suc-AAPR-pNA/suc-AAPF-pNA selectivity compared to WT-SBL for the tri-negatively charged S166C-S-CH 2

CH

2

C(COO-)

3 CMM. Conversely, the positively charged S 166C-S-CH 2

CH

2 NH3 CMM generated showed a 19-fold improvement in kcat/KM for the suc-AAPE-pNA substrate and a 54-fold improvement in suc-AAPE-pNA/suc-AAPFpNA selectivity relative to WT-SBL.

Thus, it is demonstrated herein that the addition of charge to an enzyme ;can have important effects on the activity of that enzyme against a charged substrate and that the magnitude.of that effect is a function of the amount of charge. The addition of multiple charges using site-directed mutagenesis, however, is cumbersome, typically requiring a separate mutated residue for each charge.

This difficulty is overcome herein by the use of chemically modified mutant enzymes. In a preferred embodiment, one or more cyteines are introduced into the subject enzyme via site-directed mutagenesis) and the sulfhydryl group on the cysteine provides a convenient relatively reactive thiol group that can be exploited for coupling a desired (multiply-charged) substituent to the cysteine. In a preferred embodiment, the multiplycharged substitutent of interest is provided, derivatized as a methanethiosulfonate reagent which, when reacted with the cysteine, results in the substituent of interest covalently coupled to the cysteine by a disulfide linkage Using this strategy, multiple charges WO00O 37 65 PCTUS99/30362 can be added to an enzyme with as little as one mutation, or, where the enzyme naturally possesses a suitable native cysteine, with no mutations.

C If will be noted that, in general, the location of the introduced charge is not critical. In general increasing the net positive or negative charge of the enzyme by 7 5 introducing charge essentially anywhere on the surface) will increase the ability of that OO enzyme to interact with an oppositely charged substrate. The enzyme can be routinely screened to verify that the introduced charge does not significantly diminish the activity and/or specificity of the enzyme. In certain instances it is desired to juxtapose the active site of the enzyme in question with "target substrate" in which case, it is often desirable to place 10 added charge within a subsite of the enzyme in the case of a serine protease in the Si, SI' or S2 subsite).

SI. Production of mutant enzymes for chemical modification.

A) Selection of enzymes for modification.

Virtually any enzyme can be modified (by the introduction of multiple charges) according to the methods of this invention. Such enzymes include, but are not limited to, pullulanases, proteases, cellulases, amylases, isomerases, lipases, oxidases, oxidoreductases, hydrolases, aldolases, ketolases, glycosidases, oxidoreductases, hydrolases, aldolases, ketolases, glycosidases, lyases, ligases, transferases, ligases, and the like.

Preferred enzymes for modification according to this invention include the serine hydrolases.

The serine hydrolases are a class of hydrolytic enzymes characterized by a hydrolytic enzymes that posses a catalytic triad composed ofa serine, histidine and a carboxylate amino acid (either aspartic or glutamic acid), and which catalyze the hydrolysis, and microscopic reverse reactions.thereof, of carboxylic acid derivatives including, but not restricted to, esters, peptides and amides.

Preferred serine hydrolases comprising this invention include the trypsinchymotrypsin proteases, the subtilisin proteases, and the alpha/beta hydrolases. In a particularly preferred embodiment the enzyme is protease, more preferably a subtilisin a Bacillus lentis subtilisin). Subtilisin is a serine endoprotease (MW -27,500) which is secreted in large amounts from a wide variety of Bacillus species. The protein sequence of subtilisin has been determined from at least four different species of Bacillus (see, e.g., Markland et al. (1971) pages 561-608 In: The Enzymes, ed. Boyer P. Acad Press, New 00 WO 00/37658 PCT/US99/30362 SYork, Vol. III, pp.; Nedkov et al. (1983) Hoppe-Seyler's Z. Physiol. Chem. 364: 1537-1540).

The three-dimensional crystallographic structure of subtilisin BPN' (from B.

C' amyloligoefaciens) to 2.5 A resolution has also been reported (Wright et al. (1969) Nature 221, 235-242; Drenth et al. (1972) Eur. J. Biochem. 26: 177-181. These studies indicate that although subtilisin is genetically unrelated to the mammalian serine proteases, it has a similar active site structure. The x-ray crystal structures of subtilisin containing covalently bound peptide inhibitors (Robertus, et al. (1972) Biochemistry 11: 2439-2449), product complexes (Robertus et al. (1972) Biochemistry 11: 4293-4303), and transition state analogs (Matthews et al(1975)J. Biol. Chem. 250: 7120-7126; Poulos et al. (1976) J. Biol. Chem. 251, 1097- 1103), which have been reported have also provided information regarding the active site and putative substrate binding cleft ofsubtilisin. In addition, a large number of kinetic and chemical modification studies have been reported for subtilisin (Philipp et al. (1983) Mol.

Cell. Biochem. 51:5-32; Svendsen (1976) Carlsbera Res. Comm. 41: 237-291; Markland, Id.) as well as at least one report wherein the side chain ofmethionine at residue 222 ofsubtilisin was converted by hydrogen peroxide to methionine-sulfoxide (Stauffer et al. (1965) J. Biol.

i Chem. 244: 5333-5338).

Other particularly preferred serine hydrolases for use in this invention include, but are not limited to a/p hydrolases and trypsin/chymotryspsin families of structurally serine hydrolase enzymes.

B) Selection of residues for modification.

In general, virtually any residue of the enzyme can be selected for sitedirected mutagenesis and chemical modification to introduce a multiply-charged substituent as long as the modification retains the desired level of activity of the subject enzyme.

Typically this is accomplished by making the substitution at a location that does not block critical substrate interactions or drastically alter folding/conformation of the subject enzyme.

In one preferred embodiment, residues for modification in the enzyme (e.g.

serine hydrolase) are rationally selected. Preferred sites include sites not in critical conformation determining regions and sites disposed away from the subsite(s) of the enzyme. However, in other preferred embodiments, particularly where it is desired to enhance, or otherwise alter, substrate specificity and/or activity, preferred amino acid residues selected for modification include residues expected to be important discriminatory sites within the subsites of the enzyme. Such resides are determined from mutagenesis -WO 00/3765V~ PCTIUS99/30362 experiments where the subsite residues are systematically mutagenized and the effect of such mutagenesis on binding specificity and/or enzymatic activity is determined. In addition, CK I important residues can be identified from inspection of crystal structures and/or from o predicted protein folding or protein-protein interactions determined using protein modeling software Quanta (Molecular Simulations Inc.) and Frodo academic software). Side 00 chains situated to alter interaction at subsites defined by Berger and Schecter can be selected based on the crystallographic models of the enzymes ard extrapolated to homologous enzymes if necessary if structural information on a specific enzyme is unavailable. In B.

lentus subtilisin sites 62, 156, 166, 217 and 222 are important substrate specificity determining sites. Additional related sites include position 96, 104, 107, 189 and 209 in *C subtilisin and homologous positions in related enzymes. In preferred embodiments, such residues typically lie in the S1, S1', or S2 subsites although it will be appreciated that in certain cases, alteration of residues in other subsites can also produce dramatic effects.

In one particularly preferred embodiment, where the serine hydrolase is a subtilisin-type serine hydrolase, preferred residues for mutation include, but are not limited to residues 156 and 166 in the S1 subsite, residues 217 and 222 in the S1' subsite and residue 62 in the S2 subsite Leu96, Vall04, Ile107, Phel89 and Tyr209 or residues at homologous positions within the subsites of other subtilisin-type serine proteases.

In another preferred embodiment, where the serine hydrolase is a trypsinchymotrypsin type serine hydrolase, preferred residues for mutation include, but are not limited to, Tyr94, Leu99, Glnl75, Aspl89, Serl90 and Glnl92 of trypsin or residues at homologous positions within the subsites of other trypsin-chymotrypsin-type serine proteases.

In still another preferred embodiment, where the serine hydrolase is an alpha/beta serine hydrolase, preferred residues for mutation include, but are not limited to, Trpl04, Thrl38, Leul44, Vall54, Ile189, Ala 225, Leu278 and Ile185 of Candida antartica lipase (Protein Data Bank entry I tca) or residues at homologous positions within the subsites of other alpha/beta type serine hydrolases.

Preferably the amino acids replaced in the enzyme by cysteines are selected from the group consisting of asparagine, leucine, methionine, or serine. More preferably the amino acid to be replaced is located in a subsite of the enzyme preferably the S1, S1' or S2 subsites. More preferably, in a subtilisin the amino acids to be replaced are N62, L217, M222, S156, S166, site 104, site 107 site 96 site 189(S2'), and site 209 (S -16- WO 00/37658 PCT/US99/30362 or their homologues where the numbered position corresponds to naturally occurring i subtilisin from Bacilus amyloliquefacients or to equivalent amino acid residues in other N subtilisins such as Bacillus lentus subtilisin.

SC) Introduction of cysteine.

The substitution of a cysteine for one or more native residue(s) in the enzyme serine hydrolase) can be accomplished using routine methods well known to those of ordinary skill in the art. In one preferred embodiment, the mutants described herein are most efficiently prepared by site-directed mutagenesis of the DNA encoding the wild-type enzyme of interest Bacillus lentis subtilisin). Techniques for performing site-directed mutagenesis or non-random mutagenesis are known in the art. Such methods include, but are not limited to alanine scanning mutagenesis (Cunningham and Wells (1989) Science, 244, 1081-1085), oligonucleotide-mediated mutagenesis (Adellman et al. (1983) DNA, 2, 183), cassette mutagenesis (Wells et al. (1985) Gene, 344: 315) and binding mutagenesis (Ladner et al. WO 88/06630).

In one embodiment of the present invention, the substitute amino acid residue cysteine) is introduced into the selected position by oligonucleotide-mediated mutagenesis using the polymerase chain reaction technique. In this approach, the gene encoding the desired native enzyme subtilisin) is carried by a suitable plasmid. More preferably, the plasmid is an expression vector, a plasmid from the pBR, pUC, pUB, pET or pHY4 series. The plasmid can be chosen by persons skilled in the art for convenience or as desired.

For site-directed mutagenesis, the fragment containing the selected mutation site is cleaved from the gene encoding the subject enzyme by restriction endonucleases is used as the template in a modified PCR technique (see, Higuchi et al. (1988) Nucleic Acid Res., 16, 7351-7367). For each target substitution, an oligonucleotide containing the desired mutation is used as a mismatch primer to initiate chain extension between 5' and 3 PCR flanking primers. The process includes two PCR reactions. In the first PCR, the mismatch primer and the 5' primer are used to generate a DNA fragment containing the desired base substitution. The fragment is separated from the primers by electrophoresis. After purification, it is then used as the new 5' primer in a second PCR with the 3' primer to generate the complete fragment containing the desired base substitution. After confirmation WO 00/37658 )PC US99/30362 of the mutation by sequencing, the mutant fragment is then inserted back to the position of the original fragment.

In another approach, a cassette mutagenesis method may be used to facilitate the construction and identification of the cysteine mutants of the present invention. First, the gene encoding the serine hydrolase is obtained and sequenced in whole or in part. Then the point(s) at which it is desired to make a mutation of one or more amino acids in the expressed enzyme are identified. The sequences flanking these points are evaluated for the presence of restriction sites for replacing a short segment of the gene with an oligonucleotide which when expressed will encode the desired mutants. Such restriction sites are preferably unique sites within the serine hydrolase gene so as to facilitate the replacement of the gene segment. However, any convenient restriction site which is not overly redundant in the hydrolase gene may be used, provided the gene fragments generated by restriction digestion can be reassembled in proper sequence. If restriction sites are not present at locations within a convenient distance from the selected point from 10 to 15 nucleotides), such sites are generated by substituting nucleotides in the gene in such a fashion that neither the reading frame nor the amino acids encoded are changed in the final construction. The task of S;.locating suitable flanking regions and evaluating the needed changes to arrive at two convenient restriction site sequences is made routine by the redundancy of the genetic code, a restriction enzyme map of the gene and the large number of different restriction enzymes.

If convenient flanking restriction site is available, the above method need be used only in connection with the flanking region which does not contain a site.

Mutation of the gene in order to change its sequence to conform to the desired sequence is accomplished M13 primer extension in accord with generally known methods. Once the gene is cloned, the restriction sites flanking the sequence to be mutated are digested with the cognate restriction enzymes and the end termini-complementary oligonucleotide cassette(s) are ligated into the gene. The mutagenesis is enormously simplified by this method because all of the oligonucleotides can be synthesized so as to have the same restriction sites, and no synthetic linkers are necessary to create the restriction sites.

A suitable DNA sequence computer search program simplifies the task of finding potential 5' and 3' convenient flanking sites. In preferred embodiments, any mutation introduced in creation of the restriction site(s) are silent to the final construction amino acid coding sequence. For a candidate restriction site 5' to the target codon a sequence preferably -18- 0 WO 00/37658 PCT/US99/30362 exists in the gene that contains at least all the nucleotides but for one in the recognition sequence 5' to the cut of the candidate enzyme. For example, the blunt cutting enzyme Smal (CCC/GGG) would be a good 5' candidate if a nearby 5' sequence contained NCC, CNC, or CCN. Furthermore, ifN needed to be altered to C this alteration preferably leaves the amino acid coding sequence intact. In cases where a permanent silent mutation is necessary to.

introduce a restriction site one may want to avoid the introduction of a rarely used codon. A similar situation of Smal would apply for 3' flanking sites except the sequence NGG, GNG, j or GGN must exist. The criteria for locating candidate enzymes are most relaxed for blunt cutting.enzymes and most stringent for 4 base overhang enzymes. In general many candidate sites are available.

A particularly preferred of method of introducing cysteine mutants into the enzyme of interest is illustrated with respect to the subtilisin gene from Bacillus lentus In a preferred embodiment, the gene for SBL is cloned into a bacteriophage vector M 13mp 19 vector) for mutagenesis (see, e.g. U.S. Patent 5,185,258). Oligonucleotidedirected mutagenesis is performed according to the method described by Zoller et al. (1983) Meth. Enzymol., 100: 468-500. The mutated sequence is then cloned, excised, and reintroduced into an expression plasmid plasmid GG274) in the B. subtilis host. PEG is added as a stabilizer.

The crude protein concentrate thus obtained is purified by first passing through a Sephadex T M G-25 desalting matrix with a pH 5.2 buffer 20 mM sodium acetate, 5 mM CaCl 2 to remove small molecular weight contaminants. Pooled fractions from the desalting column are then applied to a strong cation exchange column SP SepharoseTM FF) in the sodium acetate buffer described above and the SBL is eluted with a one step gradient of 0-200 mM NaCI acetate buffer, pH 5.2. Salt-free enzyme powder is obtained following dialysis of the eluent against Millipore purified water and subsequent lyophilization.

The purity of the mutant and wild-type enzymes, which are denatured by incubation with a 0.1 M HCI at 0°C for 30 minutes is ascertained by SDS-PAGE on homogeneous gels using the PhastTM system from Pharmacia, Uppsala, Sweden). The concentration of SBL is determined using the Bio-Rad (Hercules, CA) dye reagent kit which is based on the method of Bradford (1976) Anal. Biochem., 72: 248-254). Specific activity of the enzymes is determined as described below and in the examples.

SWOoo/37651'-) PCT/US99/30362 j One of ordinary skill in the art will appreciate that the protocol described O above can be routinely modified, if necessary, for use with other enzymes. Other protocols j for site-directed modification of proteins are well know to those of skill in the art and can be Sfound, for example, in U.S. Patents 5,932,419 and 5,789,166, 5,705,479, 5,635,475, 5,556,747, 5,354,670, 5,352,779, 5,284,760, and 5,071,743.

00 In addition, kits for site-directed mutagenesis are commercially available (see, e.g. TransfomerTM Site-Directed Mutagenesis Kit available from Toyobo).

D) Expression of the mutated enzyme.

In a preferred embodiment, the mutated enzyme is expressed from a 10. heterologous nucleic acid in a host cell. The expressed enzyme is then isolated and, if necessary, purified. The choice of host cell and expression vectors will to a large extent 0depend'upon the enzyme of choice and its source.

C i A useful expression vector contains an element that permits stable integration of the vector into the host cell genome or autonomous replication of the vector in a host cell independent of the genome of the host cell, and preferably one or more phenotypic markers that permit easy selection of transformed host cells. The expression vector may also include control sequences encoding a promoter, ribosome binding site, translation initiation signal, and, optionally, a repressor gene, a selectable marker or various activator genes. To permit the secretion of the expressed protein, nucleotides encoding a signal sequence may be inserted prior to the coding sequence of the gene. For expression under the direction of control sequences, a gene or cDNA encoding a mutated enzyme to be used according to the invention is operably linked to the control sequences in the proper reading frame.

Vectors containing the mutant genes obtained by site-directed mutagenesis are then used respectively to transform suitable host cells and expressed. Suitable host cells include bacteria such as E. coli or Bacillus, yeast such as S. cerevisiae, mammalian cells such as mouse fibroblast cell, or insect cells. Preferably, a bacterial expression system is used. Most preferably, the host is Bacillus. Protein expression is performed by processes well known in the art according to factors such as the selected host cell and the expression vector to culture the transformed host cell under conditions favorable for a high-level expression of the foreign plasmid.

Methods of cloning and expression of peptides are well known to those of skill in the art. See, for example, Sambrook, et al. (1989) Molecular Cloning: a Laboratory 00 i WO 00/37658 WO 0 8 PCT/US99/30362 Manual (2nd Ed., Vols. 1-3, Cold Spring Harbor Laboratory), Berger and Kimmel (1987) SMethods in Enzymology, Vol. 152: Guide to Molecular Cloning Techniques, Academic SPress, Inc. San Diego, or Ausubel et al. (1987) Current Protocols in Molecular Biology, Greene Publishing and Wiley-Interscience, New York.

As indicated above, one particularly preferred expression system is plasmid GG274 which is then expressed in a B. subtilis host.

II. Chemical modification of mutant enzyme.

A) Selection of substitutents for modifying mutated residues.

A wide variety of substitutents can be used to modify the cysteine(s) introduced into the subject enzyme serine hydrolase). As indicated above, preferred substituents are those that are multiply-charged once coupled to the target cysteine. Thus, preferred substituents, once coupled, have a net charge preferably a net charge of ±2, more preferably of and most preferably of ±4 or more.

In one preferred embodiment, the multiply charged moiety is negatively charged. Preferred multiply negatively charged moieties include, but are not limited to sulfanato, sulfinato, carboxy, phoshate, phosphite, etc. groups attached to straight chain C 2 Ci 5 alkyl, branched chain C 3 -CI5 alkyl, and to such chains containing heteroatoms such as oxygen, sulfur, nitrogen, silicon, or substituted by aliphatic, aromatic, aralkyl, haloaliphatic, haloaromatic, haloaralkyl, heterocyclic, and poycyclic combinations thereof, groups (and their C 3

-C

20 cycloalkyl, etc equivalents), to crown ethers, to cyclodextrins, to sugar moities, to monosaccharides, to oligosacharrides, to terpenoids, to alkaloids, and to sulfonatoethyl thiol, 4-carboxybutyl thiol, 3 ,5-dicarboxybenzyl thiol, 3 3 -dicarboxybutyl thiol, and 3,3,4tricarboxybutyl thiol, and the like.

In another embodiment, the multiply charged moiety (substituent) is positively charged. Preferred multiply positively moieties include, but are not limited to, primary amino, secondary amino, tertiary amino. quaternary ammonium, guanidinium, etc.

groups attached to straight chain C 2

-C

15 alkyl, branched chain C 3 -C0 5 alkyl, and to such chains containing heteroatoms such as oxygen, sulfur, nitrogen, silicon, or substituted by aliphatic, aromatic, aralkyl, haloaliphatic, haloaromatic, haloaralkyl, heterocyclic, and poycyclic combinations thereof, groups (and their C--C 2 o cycloalkyl etc equivalents), to crown ethers, to cyclodextrins, to sugar moities, to monosaccharides, to oligosacharrides, to -21- WO 0037 65 8 PCT/US99/30362 terpenoids, to alkaloids, and to aminoethyl thiol, 2-(trimethylammonium)ethyl thiol, 4,4bis(aminomethyl)-3-oxo-hexyl thiol, and 2,2-bis(aminomethyl)-3-aminopropyl thiol, and the Slike.

:O High levels'of charge can be introduced by the use of linear polymers, Z 5 branched polymers, and highly branched polymers/dendrimers. Highly charged (e.g.

00 polycation) dendrimers are well known to those of skill in the art. For example polyionic dendrimers are typically three dimensional, highly ordered oligomeric and/or polymeric compounds formed on a core molecule or designated initiator by reiterative reaction sequences adding the oligomers and/or polymers and providing an outer surface that is positively or negatively charged. These dendrimers may be prepared as disclosed in n PCT/US83/02052, and U.S. Pat. Nos. 4,507,466,4,558,120, 4,568,737,4,587,329, 4,631,337, 4,694,064, 4,713,975, 4,737,550, 4,871,779, and 4,857,599 or according to other Smethods well known to those of skill in the art. Typically, the dendrimers comprise a core molecule upon which polymers are added. The polymers may be oligomers or polymers which comprise terminal groups capable of acquiring a positive or negative charge. Suitable core molecules comprise at least two reactive residues which can be utilized for the binding of the core molecule to the oligomers and/or polymers. Examples of the reactive residues are hydroxyl, ester, amino, imino, imido, halide, carboxyl, carboxyhalide maleimide, dithiopyridyl and sulfhydryl, among others. Preferred core molecules are ammonia, tris-(2aminoethyl)amine,.lysine, ornithine, pentaerythritol and ethylenediamine, among others.

Combinations of these residues are also suitable as are other reactive residues.

Oligomers and polymers suitable for the preparation of the dendrimer polycations of the invention include, but are not limited to polyamidoamines derived from the reaction of an alkyl ester of an a,P-ethylenically unsaturated carboxylic acid or an .a,Pethylenically unsaturated amide and an alkylene polyamine or a polyalkylene polyamine, among others. Preferred are methyl acrylate and ethylenediamine.

Preferred multiply-charged polymers and/or dendrimers for use in the methods of this invention include, but are not limited to essentially any charged polymer and/or dendrimer. Preferred dendrimers include, but are not limited to glycodendrimers.

Glycodendrimers are well known to those of skill in the art (see, Parameswaran (1990) Org. Prep. Proc. Intl., 22: 119-121, Green (1970) Meth. Enzym., 18A: 418-424, and U.S.

Patent 2,489,237).

00 0 WO 0037658 PCT/US99/30362 SWhile in certain embodiments, the introduced multiply charged moieties are either positively charged or negatively charged, this invention need not be so restricted.

-C Thus, for example, where it is desired to introduce a dipole moment to the subject enzyme, multiply-charged positive moieties can be introduced at one location in the enzyme and multiply-charged negative moieties can be introduced at another location within the enzyme.

B) Coupling substituents to the cvsteine.

The R group on cysteines provides a convenient relatively reactive thiol group that can be exploited for coupling a desired multiply-charged substituent to the cysteine. In a preferred embodiment, the substitutentof interest is provided, derivatized as a methanethiosulfonate reagent which, when reacted with the cysteine, results in the substituent of interest covalently coupled to the cysteine by a disulfide linkage In a preferred embodiment, chemical modification with the methanethiosulfonate reagent(s) is carried out as described by Berglund et al. (1997) J. Am.

Chem. Soc., 119: 5265-5255 and DeSantis et al. (1998) Biochemistry, 37: 5968-5973.

Briefly, 200 pL of a 1 M solution of the methanethiosulfonate (MTS) reagent is added to a solution (5-10 mg/mL, 3.5 mL) of the cysteine mutant in 70 mM CHES, 5mM MES, 2 mM CaCI 2 pH 9.5. The MTS reagent is added in two portions over 30 minutes. Reaction mixtures are kept at 20 0 C with continuous end-over-end mixing. Reactions are monitored by following the specific activity with suc-AAPF-pNA) and by tests for residual free thiol with Ellman's reagent). Once the reaction is complete, the reaction mixture is loaded on a SephadexTM PD-10 G25 column with 5 mM MES and 2 mM CaC12, pH 6.5. The protein fraction is then dialyzed against 1 mM CaC12 and the dialysate is lyophilized.

In certain instances, where the substituent that is to be coupled to the cysteine, Sbears reactive groups the reactive groups may be derivatized with appropriate blocking/protecting groups to prevent undesired reactions during the coupling. Similarly, where the serine hydrolase contains one or more cysteines that are not to be derivatized, the thiol group(s) on these cysteines may be derivatized with appropriate protecting groups (e.g.

benzyl, trityl, tert-butyl, MOM, acetyl, thiocarbonate, thiocarbamate, and others). The use of blocking/protecting groups is well know to those of skill in the art (see, e.g., Protective Groups in Organic Synthesis" Theodora W. Greene and Peter G. M. Wuts Third Edition, Wiley-Interscience, Toronto, (1999), pp 454-493.) -23- WO 00/37658 )PCTIUS99/30362 7t III. Screening chemically modified mutants for desired activity.

0 SThe multiply-charged chemically modified mutant enzymes of this invention are typically screened for the activity or activities of interest. Such activities include, but are O not limited to, protease activity, lipase activity, cellulase activity, amylase activity, and the Z 5 like. In a preferred embodiment the activity is measured against one or more test substrates, 00 e.g. as illustrated in the examples. Thus, for example, the multiply charged mutant proteases can be assayed against synthetic substrates such the suc-AAPA-pNA substrate and suc- M AAPF-pNA substrates illustrated in the examples. Other suitable synthetic substrates for measuring enzymatic activity are well known to those of skill in the art (see, Colowick i 10 (1970) Proteolytic Enzymes, Meth. Enzymology, Vol. 19 Academic Press, New York and C London; Alan Fersht (1985) Enzyme Structure and Mechanism, 2nd Edition Alan Fersht, W.

H. Freeman and Co., New York 1985).

SOther assays can be utilized that more closely approximate the commercial applications contemplated for the modified enzymes of this invention. Thus, for example,hydrolases, are often a component of many of the detergent compositions that are currently on the market. One challenge facing a detergent manufacturer is the identification of new and improved enzymes and detergent compositions. Several factors can effect the determination of the "improvement" of a new enzyme a multiply-charged chemically modified enzyme of this invention). Such factors include, but are not limited to the enzyme, the wash conditions, and the detergent composition that the enzyme is mixed with. For example, an enzyme that works well in one detergent composition may not perform as well in another. Similarly an enzyme and/or detergent composition may perform well under one set of wash conditions, e.g. Japanese (characterized by raw cottons and low surfactant levels i resulting in positively charged textile surfaces), but not another set of wash conditions (e.g.

North American characterized by high surfactant levels resulting in negatively charged textile surfaces).

Thus, in one embodiment this invention provides methods of assaying for a preferred enzyme and/or preferred enzyme and/or detergent composition. These methods are particularly well suited to evaluating the performance of multiply charged chemically modified mutant hydrolases of this invention. The methods preferably involve providing a swatch that typically includes a piece of material a fabric such as cotton, wool, hemp, paper, and synthetic materials polyester), and mixtures of natural and synthetic fibers) 00 WO 00/37658 PCT/US99/30362 and a stain blood, milk, ink, grass, gravy, chocolate, egg, cheese, clay, pigment, oil, and combinations thereof). One particularly preferred stain is a blood/milk/ink (BMI) stain. The M stain is fixed to the material and a smaller swatch can, optionally, be removed from the Sswatch as a test or control). An enzyme is applied to he swatch or smaller swatch and they are incubated together with or without agitation and with or without subsequent rinse.

The method can further include measuring the degree of removal of the stain from the material. In one embodiment the method can involve agitating the smaller swatch and enzyme during incubation. The enzyme can be applied to the swatch or smaller swatch in combination with one or more a detergent ingredients.

The degree of removal of the stain can then be evaluated, e.g. by visual inspection or by measuring the absorbance and/or fluorescence of, for example, ink, labeled blood, or labeled milk in a supernatant after the swatch has been incubated with an enzyme and/or detergent composition.

One aspect of this invention includes a method of agitating a microtiter plate (or other test container) to a sufficient degree to assure complete and efficient incubation of the enzyme with the smaller swatch. The methods preferably involves applying a plate sealer to the top of a microtiter plate and then clamping another lid on top of the plate sealer.

The assay ;methods of this invention preferably involve the use of a treatment protocol that allows one to control of the degree of stain fixation. The use of fixed swatches leads to a dramatic improvement in the signal-to-noise ration in the wash assays. Thus, for example, untreated BMI watches washed in detergent without bleach typically release a large portion of ink, even without the help of a protease. Adding a protease leads to a small increase in ink release, but this is difficult to quantify over the large background. By :i controlling the degree of fixation, according to the methods of this invention, it is possible to produce swatches that release varying amounts of ink when washed in the absence of protease. Thus, the background "ink" signal can be reduced leaving an assay more sensitive to the effects of the protease. In addition, by varying the degree of fixation, one can generate stains that give optimum results under the various cleaning conditions.

The stain can be fixed to the material in a number of ways. For example, the swatch can be incubated with a cross-linking agent to fix the stain. The degree of fixation can be effected by, for example, increasing or decreasing the incubation time, varying the temperature at which the incubation takes place, and/or varying the concentration of the cross-linking chemical and/or stain agent(s).

W 0037658) PCTUS99/30362 Alternatively the assay can be optimized by using swatches having stains of known "strength". Swatches having strains of known "strength' on various types of material C1 are commercially available (EMPA, St. Gallen, Switzerland; wfk Testgewebe GmbH, SKrefld Germany; or Center for Test Materials, Viaardingen, The Netherlands) and/or can be made by the practitioner (Morris and Prato (1982) Textile Research Journal, 52(4): 280- 00 286). One particularly preferred swatch is BMI on a cotton swatch.

In one preferred embodiment of this invention, a BMI stain can be fixed to cotton with 0.0003 to 0.3% hydrogen peroxide. Other combinations include, but are not limited to, grass or spinach fixed with glutaraldehyde, gelatin and Coomassie stain fixed with 10 glutaraldehyde, or chocolate milk and soot fixed with glutaraldehyde.

Trinitrobenzenesulfonic acid (TNBS) can be used to quantify the concentration of amino groups in the wash liquor. This can serve as a measure of the amount 0 of protein that was removed from the swatch (see, Cayot and Tainturier (1997) Anal.

Biochem., 249: 184-0200). However, if a detergent or an enzyme sample leads to the formation of unusually small peptide fragments from the presence ofpeptidases in the sample) then one will obtain a larger TNBS signal, more noise.

The present invention provides another and a better way to measure wash performance that is based on ink release. For example, blood/milk/ink swatches (BMI) are used. Proteolysis of protein in the swatches leads to the release of ink particles which can be quantified by measuring the absorbance of the wash liquor. The absorbance can be measured at any wavelength between 350 and 800 nm. IN a preferred embodiment, the wavelength is measured at 410 nm or 620 nm.

In a preferred embodiment of the invention, a BMI stain is fixed to cotton by applying 0.3% hydrogen peroxide to the BMI/cotton swatch for 30 minutes at 25 0 C or by applying 0.03% hydrogen peroxide for 30 minutes at 60 0 C. Smaller swatches (e.g.

approximately 0.25 inches) are cut from the BMI/cotton swatch and placed in the wells of a 96 well microtiter plate. Into each well, a known mixture of a detergent composition and an enzyme such as a multiply charged chemically modified mutant of this invention, is placed.

After placing an adhesive plate sealer onto the top of the microtiter plate, the microtiter plate is clamped to an aluminum plate and agitated for 10-200 minutes. At the end of this time, the supematants are transferred to wells in a new microtiter plate and the absorbance of the ink at 620 nm is measured.

-26- 0 0 o, WO 00/37658 PCT/US99/30362 The performance of variant proteases (produced, for example, according to the disclosure of UUSSN 322,678) in the MSA using TNBS and ink release detection were compared. Several of these samples showed inflated wash performance when TNBS detection was used (probably due to peptidase contamination) whereas all samples result in indistinguishable signals when the absorbance of the wash liquor was measured.

The dose response of variant proteases in granular detergent under Japanese wash conditions can be examined. The absorbance of the wash liquor can be determined.

Subsequently the swatches are rinsed and dried and the reflectance is measured with a flat bed scanner. AN excellent correlation between both detection methods is observed which demonstrates that the absorbance of the wash liquor is a true measure of the cleaning performance. Measuring ink release is simpler, more precise, and better suited to highthroughput screening methodologies than measuring the reflectance of swatches.

To increase the throughput and make the assay suitable for screening large numbers of samples, it can be adapted to 96 well microtiter plates by using smaller (e.g.

l/4")swatches. However the measurements can result in relatively large standard deviations. It has been observed that the signal is dependent on the orientation of the swatches in the wells (horizontal versus vertical) which indicates that mixing was insufficient. To address this issue, a plate holder in which the microtiter plate is sandwiched between two outer plates aluminum plates) can be constructed. The device can be mounted in a commercial incubator shaker. Setting the shaker to 400 rpm resulted in very efficient mixing, while leakage or cross-contamination between the wells was effectively prevented by the holder.

It will be appreciated that the forgoing assays are not limited to textile washing. Similar assays can be performed for essentially any other cleaning application.

Thus, for example, the present invention can be used to determine a preferred enzyme and/or detergent composition for dish washing, for example, using a BMI stain on a suitable substrate such as cloth, plastic, glass, or ceramic. It will be appreciated that the foregoing protocol is exemplary and not limiting and numerous modifications and variants can be performed with only routine experimentation by one of ordinary skill in the art.

Production of multiply-charged chemically modified mutant enzymes and screening for particular activities of such modified enzymes is amenable to high throughput methodologies. Typically such methodologies utilize robotics to automate and speed the production and screening of large numbers of compounds. In efficient high throughput WO 00/37658 PCT/US99/30362 screening system, typically hundreds of thousands of reactants/reactions can be screened in a Sfew days with only routine operator involvement. High throughput screening systems are commercially available (see, Zymark Corp., Hopkinton, MA; Air Technical Industries, O Mentor, OH; Beckman Instruments, Inc. Fullerton, CA; Precision Systems, Inc., Natick, Z 5 MA, etc.). These systems typically automate entire procedures including all sample and 00 reagent pipetting, liquid dispensing, timed incubations, and final readings of the microplate in detector(s) appropriate for the assay. These configurable systems provide high throughput and rapid start up as well as a high degree of flexibility and customization. The Smanufacturers of such systems provide detailed protocols the various high throughput. Thus, for example, Zymark Corp. provides technical bulletins describing screening systems for C' detecting the modulation of gene transcription, ligand binding, and the like.

SIV. Uses of the CMMs of this Invention.

The multiply-charged chemically modified mutant enzymes of this invention can be used in a wide variety of applications. The applications can be the same as those in which "conventional" enzymes of the same type are used, the multiply charged variants of this invention offering improved performance stability, specificity, activity, etc.). In preferred embodiments, the enzymes of this invention will be used where enhanced performance on charged substrates is desired.

Thus, for example, the modified hydrolases of this invention proteases, cellulases, amylases, laccases, lipases, etc.) may be formulated into known powdered and liquid detergents. The particular multiple-charged variant will depend on the anticipated wash conditions. Thus, for example, in Japan where wash conditions typically involve naked cotton and low surfactant concentrations, the cloth surface tends to accumulate a positive charge and negative multiply charged modified enzymes are desirable. Conversely in Europe where high surfactant use results in negatively charged surfaces, positive multiply charged modified enzymes are desirable.

Typical detergent compositions will have contain the multiply charged enzymes at levels of about 0.0 Ito about preferably about 0.1% to about by weight. These detergent cleaning compositions or additives can also include other enzymes components such as builders and stabilizers, etc..

In particularly preferred embodiments, the multiply charged proteases, more preferably multiply charged subtilisins are used in formulating various detergent -28- W000/37658 PCT/US99/30362 compositions. A number of known compounds are suitable surfactants useful in such detergent compositions. These include nonionic, anionic, cationic, anionic, or zwitterionic detergents (see, U.S. Patents 4,404,128, and 4.261,868). A suitable detergent formulation is that described in example 7 of U.S. Patent 5,204,015. The modified enzymes of this invention may provide improved was performance in a detergent composition (as compared to previously known additives). Improves wash performance typically refers to increased cleaning of certain modified enzyme-sensitive stains such as grass or blood, as determined by a standard evaluation procedure light reflectance) after a standard wash cycle.

The art is familiar with the different formulations that can be used as cleaning compositions. In addition to typical compositions, it is readily understood that the modified enzymes of the present invention may be used for any purpose that the native or wild-type enzymes are used. Thus, these modified enzymes can be used, for example, in bar or liquid soap applications, dish care formulations, contact lens cleaning solutions or products, peptide synthesis, feed applications such as feed additives or preparation of feed additives, waste treatment, textile application such as the treatment of fabrics, and as fusion-cleavage enzymes in protein production.

In another preferred embodiment, the modified enzymes of this invention are used in a method of treating a textile. The methods involve contacting a multiply charged chemically modified mutant enzyme of this invention with a textile under conditions effective to produce a textile resistant to certain enzyme-sensitive stains grass or blood stains). The method can be used to treat, for example, silk or wool. Enzyme treatments of such fabrics are know to those of skill in the art and are described for example in Research Disclosure 216,034, European Patent application No: 134,267, U.S. patent 4,533,359, and European Patent application 3244,259.

In still another embodiment, the modified enzymes of this invention are used in the preparation of an animal feed, for example, a cereal-based feed. The enzyme can be incorporated into essentially any cereal feed, e.g. a cereal comprising one or more of wheat, barley, maize, sorghum, rye, oats, triticale, and rice. Although the cereal component of a cereal-based feed constitutes a source of protein, it is usually necessary to include species of supplementary protein in the feed such as those derived form fish meal, meat, or vegetables.

Sources of vegetable proteins include, but are not limited to soybeans, rape seeds, canola, soybean meal, rapeseed meal, and canola meal.

-29w o o, 0 WO 00/37658 i) PCT/US99/30362 The inclusion of a multiply charged modified enzyme in an animal feed can 0 enable the crude protein value and/or the digestibility and/or the amino acid content of the feed to be increased. This permits a reduction in the amounts of alternative protein sources O and/or amino acid supplements that are added to the feed.

The foregoing description of uses for the modified mutant enzymes of this 00 _invention is illustrative and not intended to create any special use limitation. One will appreciate that the uses of the enzymes of this invention are myriad and not to be confined to Sthe uses enumerated herein.

V. Kits and products containing chemically modified mutants.

In still another embodiment, this invention provides kits for synthesizing and/or screening multiply charged modified mutant enzymes of this invention. Such kits 0 preferably include one or more mutant serine hydrolases having one or more amino acid residues substituted with a cysteine as described herein and/or include one or more multiplycharged methane sulfonate reagents as described herein that can be used to derivatized the mutant serine hydrolase. Such kits may additionally include one or more reagents and/or apparatus for performing such derivitizations.

In addition, the kits can include appropriate substrates and/or reactants for screening the chemically modified mutant enzyme for one or more desired activities as described herein.

In another embodiment this invention provides kits for the use of the modified mutant enzymes described herein. Such kits preferably contain one or more containers containing one or more of the chemically modified mutant serine hydrolases as described herein. Such kits can also include appropriate reagents and/or substrates to use the modified enzymes in one or more of the reactions described herein.

In addition, the kits may include instructional materials containing directions I protocols) preparation of the multiply charged enzymes of this invention and/or for their use. Thus, for example, in one preferred embodiment, the instructional materials provide protocols derivatizing the mutant enzyme with one or more of the multiply charged methane sulfonate reagents described herein. In another embodiment, the instructional materials may provide protocols describing the use of the modified enzyme as an additive to a cleaning agent a laundry detergent). While the instructional materials typically comprise written or printed materials they are not limited to such. Any medium capable of storing such 00 WO 00/37658 PCT/US99/30362 Sinstructions and communicating them to an end user is contemplated by this invention. Such Smedia include, but are not limited to electronic storage media magnetic discs, tapes, cartridges, chips), optical media CD ROM), and the like. Such media may include Saddresses to internet sites that provide such instructional materials.

C 5

EXAMPLES

The following examples are offered to illustrate, but not to limit the claimed invention.

Example 1 The Controlled Introduction of Multiple Negative Charge at Single Amino Acid Sites in Subtilisin Bacillus lentus SUsing the X-ray structure of SBL (the coordinates of SBL have been deposited at the Protein Databank at Brookhaven under the code IJEA) (Figure 1) as our guide, four sites were chosen for mutation because of their seminal positions in the active site. Two of these, N62 (subtilisin BPN' numbering) and L217, occupy positions that are equidistant from S221 of the catalytic triad, in the S 2 (Nomenclature of Schechter and Berger (1967) Biochem.

SiBiophys. Res. Commun. 27: 157-162) and Si' pockets respectively. The other two sites, S156 and S166, are located at the base ofthe Si pocket and their side chains are directed towards SBL's surface and catalytic triad respectively. The MTS reagents 1'-e were selected to modify these positions.

Results And Discussion Synthesis of Carboxvalkyl Methanethiosulfonates lb-e Previous work (Kenyon and Bruice (1977) Methods Enzymol. 47: 407-430; Berglund et al. (1997) J. Am. Chem. Soc. 119: 5265-5266) has demonstrated that, of the methods available (Douglass and Farah (1959) J. Org. Chem. 24: 973-975; Levitt and Levitt (1972) J. Org. Chem. 37: 332-334; Weidner and Block (19.72) J. Med. Chem. 15: 564-567; Palumbo and Caputo (1981) Synthesis 888-890; Block and Zhao (1992) J. Org. Chem. 57: 5815-5817; Billard et al. (1996) J. Org. Chem. 61: 7545-7550), direct nucleophilic displacement of a primary alkyl bromide by methanethiosulfonate ion provides the most efficient method for the preparation of alkyl methanethiosulfonates. This general method was therefore adopted as the basis for the preparation of all of lb-e. The aliphatic -31- WO00/3765f) PCTtUS99/3D362 monocarboxylate MTS lb Patent 4,879,249) was prepared from acid and NaSS02CH 3 in 80% yield.

The aromatic dicarboxylate MTS Ic was prepared from dicarboxylic acid via a precursor benzylic bromide 3 as shown in Scheme 2 (Figure 2).

Treatment of 3 with NaSSO 2

CH

3 gave the protected aromatic MTS 4 in 60% yield.

Hydrolysis of 4 with TFA gave Ic as a white solid (91% yield, 44% overall yield from 2).

The aliphatic di- and tricarboxylates ld,e were prepared from Meldrum's acid using 1,2-dibromoethane to introduce a brominated linker group as shown in Scheme 3 (Figure The low pKa of a (Arett et al. (1984) J. Am. Chem. Soc. 106:6759-6766; Arnett and Harrelson (1987) J. Am. Chem. Soc. 109: 809-812) allowed the use of mildly basic conditions compatible with this choice of linker. For the sake of simplicity, we chose methyl Meldrum's acid (5b) as a starting material for Id in which one alkylation site is blocked as a methyl group. Direct alkylation of 5a with 1,2-dibromoethane led only to the formation of a spirocyclopropane derivative, the product of intramolecular cyclization, in low yield. For a recent analysis of the high propensity of the anion of 5a to form cyclic products with a.,o-dihalides see Ridvan (1997) J. Tetrahedron 53: 14793-14806. The synthesis of le utilized this position to introduce a third carboxylate moiety.

Alkylation of 5b with 1,2-dibromoethane afforded bromide 6b in 71% yield.

Treatment of 6b with NaSS0 2

CH

3 in DMF at 50' C led to the displacement of the remaining bromide and resulted in the formation of protected dicarboxylate MTS 7b. Hydrolysis of 7b using acidic ion exchange resin allowed the successful formation of the aliphatic diacidic MTS ld (79% yield, 37% overall yield from The synthesis of the triacidic reagent le required the construction of a protected tricarboxylate 5c before elaboration. Alkylation of Meldrum's acid (5a) with tertbutyl bromoacetate allowed the formation of Sc with moderate selectivity in 59% yield.

Elaboration of tricarboxylate 5c was carried out using 1,2-dibromoethane and then NaSSO 2

CH

3 in an analogous manner to that used for the synthesis of Id and gave protected tricarboxylate MTS 7c in 57% yield over 2 steps. Complete deprotection of 7c using

CF

3 COOD in D 2 0 was followed by 'H NMR, and resulted in the formation of target le (70% yield, 23% overall yield from WO 00/37658 PCT/US99/30362 Preparation of Chemically Modified Mutants (CMMs)

S

M

TS reagents la-e were used to modify the chosen SBL cysteine mutants, N62C, S 156C, S 166C and L217C under conditions described previously (Berglund et al.

(1997) J. Am. Chem. Soc. 119: 5265-5266; Stabile et al. (1996) Bioorg. Med: Chem. Lett.6: 2501-2512; DeSantis et al. (1998) Biochem. 37: 5968-5973). These reactions proceeded rapidly and quantitatively, as judged by the monitoring of changes in specific activity and by titration of free thiols with Ellman's reagent (Ellman et al. (1961) Biochem. Pharmacol. 7: 88-95), respectively. The structure of the charged CMMs was confirmed by ES-MS. Nonreducing native PAGE was used to determine the purity of all the enzymes, which appeared 10 as single bands. Consistent with the introduction of negative charge, each of the CMMs Sshowed retarded mobility in the direction of the cathode relative to WT. The active enzyme concentration of the resulting CMM solutions was determined by active site titration with atoluenesulfonyl fluoride (PMSF) using a fluoride ion-sensitive electrode (Hsia et al. (196) J.

Anal. Biochem. 242: 221-227).

Kinetic Effects of Site Specific Modification The effects of modification upon SBL were assessed by the determination of ka and KM for the hydrolysis of succinyl-AAPF-p-nitroanilide (Suc-AAPF-pNA). Our usual assay pH of 8.6 ensured complete ionization of the unnatural amino acid side-chains introduced. The kinetic parameters of the 20 CMMs generated are compared with those of WT and unmodified mutants in Table I and Figure 4.

Table 1. Kinetic Parametersa for Modified Enzymes Entry Enzyme Pocket R Level of ke, Ku (mM) kaJ/KM Charge (s-'mM 1 1 SBL-WT 153 4 0.73 ±0.05 209 2 N62C S 2 H 0 174 9 1.90 0.20 92 11 3 a 1 119 ±4 0.93 0.07 128 11 4 b 1 106±2 1.01 ±0.05 105 6 c 2 113 7 1.00±0.10 113 13 6 d 2 .90±4 1.47±0.14 61±6 7 e 3 129 3 1.46 ±0.06 88 4 8 L217C Si' H 0 41 1 0.80 i 0.04 51 3 9 a 1 89 6 1.80 0.20 50 6 b 1 54 1 1.03 0.06 52 3 -33- WO 0037658 )PCTUS99/30362 11 c 2 69 2 .0.81 ±0.06 85 7 12 d 2 63±2 1.65±0.11 38±3 C 13 e 3 55 2 1.48 0.08 37± 3 14 S156C St H 0 125 4 0.85 0.06 147±11 0 15 a 1 87 2 1.20 0.07 73± Z 16 b 1 76 ±1 1.08 0.04 70 3 00 17 c 2 61 1 1.39 0.10 44 3 18 d 2 53 ±1 1.67 ±0.06 32 1 19 e 3 74± 2 1.87 0.08 39 2 S166C H 0 42 1 0.50 0.05 84 9 21 a 1 25± 1 1.34 ±0.08 19 1 22 b 1 48 Ib 1.52 0.06 31 I b H: 23 c 2 47 3 1.60 0.20 29 4 24 d 2 67 ±2 2.26 ±0.10 30 2 e 3 76±2 2.46±0.11 31±2 Michaelis-Menten constants were measured at 25 0 C according to the initial rates method in 0.1 M Tris-HCl buffer at pH 8.6,0.005% Tween 80, 1% DMSO, Suc-AAPF-pNA as the substrate.

b Based on total protein concentration. Certain inconsistencies between active enzyme concentration as determined by active site titration with PMSF (Hsia etal. (1996)J. Anal. Blochem. 242: 221-227) and total protein concentration have been reported for negatively charged mutants of SBL. These are characterized by sluggish fluoride ion concentration bursts and high rates of background PMSF hydrolysis. Active enzyme concentration values for S166C-b were low and gave rise to an anomalous value for kcat (270 5 s-1).

Consequently the value shown was calculated using total protein concentration as determined by absorbance at 280 nm (e280 23000 M-l cm-1) (Gren et al. (1990) Eur. J. Biochem. 194 :897-901). The KM value determined (1.52 0.06 mM) is not concentration dependent Position 62 in the S2 pocket is the most tolerant of mutation and modification (Figure 4A) and mutation to cysteine reduces kcal,Ku by only a factor of 2 (Table 1, Entry 2).

Upon modification, activity is partially restored and values ofkca,/KM for N62C-a,b,c are raised approximately 1.5-fold relative to N62C (Table 1, Entries Modifications with aliphatic di- and tri-carboxylate MTS reagents ld,e elicit further drops in kca,/KM, with i N62C-d being 3.5-fold lower than WT. However, despite the increased charge, this kca,/KM drop is less marked for N62C-e.

The deleterious effect of negative charges in the S2 pocket upon kc, is apparent in the 1.3-fold decrease observed for N62C-a (Table 1, Entry 3) relative to WT.

However, as the level of negative charge increases, kcat values do not decrease further to any significant extent. In fact, of all the CMMs, the kca, level of N62C-e (129 Table 1, Entry 7) is the highest. In contrast, KM values increase continually with each additional charge, reaching values for N62C-d (Table 1, Entry 6) and N62C-e (Table 1, Entry 7) that are 2-fold higher than WT.

77 SWO 0.0/37658 PCT/US99/30362 The effects of mutation at position 217 in the SI' pocket (Figure 4B) are intrinsically more dramatic than at all three other sites since the value of k.c/Ku for L2i7C is 4-fold lower than WT. The introduction of a single negative charge only affects kcKu a little and leads to near-identical kcaI/K values for L217C-a,b (Table 1, Entries 9,10). As negative charge, increases further, two opposite trends are observed, with the kcai/Ku value for aromatic L217C-c being raised 1.6-fold relative to L217C-a,b, while those for aliphatic L217C-d,e are reduced by 1.4-fold.

These slight kca,/KM changes seen at position 217 are the result of larger, but i counteracting changes in each of kt and KM. For example, while L217C-a has the highest value of kcar, it also has the highest KM value (both 2.2-fold higher than L217C). As at position 62, when the level of negative charge increases, from L217C-a to L217C-e, kcai values decrease only slightly and remain 1.3- to 1.7-fold higher than L217C (Table 1, Entries 10-13). KM values increase unevenly to 2-fold higher than WT. Interestingly, the underlying cause of the out-of-line kca/Ku of L217C-c is an unusually low KM (0.81 mM, Entry 11), which may be a consequence of complementary aromatic interactions between the substrate and the phenyl ring of side chain c.

The effects of mutation and modification at positions 156 and 166 in the S1 pocket are shown in Figures 4C and 4D. Mutation at position 156 to cysteine causes a 1.4fold drop in ka/KM (S156C, Table 1, Entry 14). From S 156C-a to S 156C-d kcaJK decreases monotonically to a value that is 6-fold lower than WT. The additional negative charge present in S 156C-e partially restores this value, to only 5.4-fold lower than WT.

Mutation and modification at position 166 leads to the least active negatively charged CMMs with kcailKus 6- to 11 -fold lower than WT. This partly reflects the intrinsically lower kca/Ku value of the unmodified mutant S 166C, which is already I 25 lower than WT. However, the presence of the sulfonatoethyl side chain in S166C-a causes a dramatic drop to a value that is 11-fold lower than WT. kc,,/KMis increased 1.5-fold for S166C-b and remains steady as the level of negative charge increases from S 66C-c to S166C-e.

The kca, values for S156C and S166C CMMs are similar to those found for L217C CMMs, typically 2 to 2.5-fold lower than WT. As at positions 62 and 217, the detrimental effect of a single negative charge on ki is not amplified by the introduction of additional negative charges. In fact, kcEa values for S166C CMMs increase steadily from 6- WO 00/376581 PCT/US99/30362 fold lower than WT for S166C-a (Table 1, Entry 21) to 2-fold lower than WT for S166C-e 0(Table 1, Entry I The KM values for both S 156C and S 166C CMMs increase steadily with O increasing negative charge and are largest for S166C-e (KM 2.46 mM, 3.5-fold higher than WT, Table 1, Entry 25). Consistent with its surface-exposed nature these effects are less 00 I pronounced at position 156 with KM increasing to only 2.5-fold higher than WT for S156C-e (Table 1, Entry 19).

n i Kinetic Effects of Negative Charge i To separate the contribution of electrostatics from steric effects, a comparison of these charged CMMs with those containing sterically similar uncharged side chains was made. (The following kinetic parameters for the previously prepared (Berglund et al. (1997) J. Am. Chem. Soc. 119: 5265-5266; DeSantis et al. (1998) Biochem. 37: 5968-5973) near isosteric CMMs were determined under the conditions described in the experimental section; N62C-S-ethyl kcai 104 2 KM 0.55 0.04 mM, kca/Ku 189 14 N62C-S-benzyl 129 3 KM 0.34 0.03 mM, kc,/KM 379 37 N62C-S-n-pentyl kca, 184 KM 0.75 0.05 mM, kca/KM 245 18 L217C-S-n-pentyl 87 3 KM 0.52 0.05 mM, kca/Ku 167 17 s-mM'; S156C-S-benzyl kc, 72 2 Ku 0.59 0.05 mM, kcCJKM 122 11 S166C-S-ethyl kcat 11.8 0.5 KM 0.76 0.08 mM, kc/KM 15.5 1.8 s'mM'; S166C-S-benzyl kc, 23.1 0.5 KM 1.17 0.06 mM, kca/KM 19.7 1.1 s" For example, N62C-a was compared with N62C-S-ethyl, N62C-b,d,e were compared with N62C-S-n-pentyl and N62C-c was compared with N62C-S-benzyl. Figure illustrates the results of introducing charge to these near-isosteric systems. This provides an estimate of the effect of negative charge upon the kinetics of SBL when corrected for underlying steric and hydrophobic effects.

Two differing trends emerge from Figure 5. At positions 62, 217 and 156, the electrostatic contribution of each of side chains a-e is detrimental to keIKu. The reductions caused are similar for each side chain, vary little from site to site and increase with the level of negative charge introduced. These reduced kca/KM values resulting from the introduction of negative charge are consistent with earlier findings (Ballinger et al. (1995) Biochem. 34: 13312-13319; Berglund et al. (1997) J. Am. Chem. Soc. 119: 5265-5266; Stabile et al. (1996) Bioorg. Med. Chem. Lett.6: 2501-2512; DeSantis et al. (1998) Biochem. 37: 5968-5973.).

-36- WO 00/37658 PCT/US99/30362 Such effects may be attributed, in part, to destabilization of the tetrahedral oxyanion intermediate that is formed in the rate limiting step of catalysis (Jackson and Fersht (1993) Biochem. 32: 13919-13916).

In contrast, the introduction of negative charge at position 166 partially restores some of the activity lost through the introduction of near-isosteric uncharged side chains. Therefore the drastically lowered kJKMs of CMMs S 166C-a-e relative to WT are, in fact, a result of steric or hydrophobic effects. Mutation analysis of subtilisin BPN' has shown that kCa,/KM decreases dramatically when the optimal binding volume of the Sipocket is exceeded (Estell et al. (1986) Science 233:659-663). The effect of introducing even small groups at position 166 of SBL (this space is more limited in SBL than in subtilisin BPN' as the peptide backbone that makes up the wall of the S pocket contains four less amino acid residues) is to fill the S-pocket and this dramatically decreases k,/Kau. For example, uncharged CMM S166C-S-ethyl has a kca/K 13.5-fold lower than WT. Molecular mechanics analysis of S 66C CMMs has shown that charged side chains introduced at position 166 may orientate themselves towards external solvent (DeSantis e al. (1998) Biochem. 37: 5968-5973). This serves to reduce the volume of the SI pocket that is occupied by the side chain. The existence of such an orientation for S166C-a-e, which is lacking in uncharged CMM counterparts, might, in part, explain the beneficial effects of introducing charge. As a result, charged CMM S166C-S-EtSO3-, side chain a, has a kcar/Ku only 11-fold 20 lower than WT.

j Conclusions In summary, we have devised short and efficient synthetic routes to three novel multiply charged methanethiosulfonates, Ic, d and e. Such compounds, as well as being of interest in our approach to the controlled tailoring of enzyme activity, may prove useful in the study of ion channels. The use of MTS reagents in techniques such as the substituted-cysteine accessibility method (SCAM) (Akabas et al. (1992) Science 258: 307- 310; Akabas et al. (1994) Neuron 13: 919-927; Akabas et al. (1994) J. Biol. Chem. 269: 14865-14868) has allowed aspects of membrane ion channel topology and conformation to be determined. In particular the use of charged MTS reagents has given an invaluable insight into ion specificity (Cheung and Akabas (1997) J. Gen. Physiol. 109: 289-299) and mechanism of action (Stauffer and Karlin (1994) Biochem. 33: 6840-6849; Yang et al.

(1996) Neuron 16: 113-122; Holmgren et al. (1996) Neuropharmacol. 35: 797-804; Huynh -37- WOo/376) i PCT/US99/30362 et al. (1997) J. Gen. Physiol. 110: 229-242; Rassendren et al. (1997) EMBOJ. 16: 3446- 3454).

Using our established methodology, we selectively modified the cysteine thiols of SBL mutants, N62C, S156C, S166C, and L217C, with these reagents. Without exception, mutation and modification at all four sites led to reduced catalytic efficiency in the hydrolysis of Suc-AAPF-pNA. However these reductions do not exceed 11-fold relative to WT and the lowest kl values determined were only 6-fold reduced. This reduced efficiency is manifested largely through decreased binding interactions, decreased KM values, that increase with the level of charge introduced. In contrast, kca, values corresponding to the introduction of multiple charge are similar to, if not higher than, those for single charge. Comparison with near-isosteric uncharged CMMs revealed that electrostatic effects are paramount at positions 62, 217 and 156. However at position 166 steric effects dominate and the introduction of negative charge is, in fact, beneficial.

Experimental Mutants of Subtilisin Bacillus lentus (SBL) were generated, and WT and mutant enzymes purified as described previously (Stabile et al. (1996) Bioorg. Med. Chem.

Lett.6: 2501-2512; DeSantis et al. (1998) Biochem. 37: 5968-5973). NaSSO 2

CH

3 (Kenyon and Bruice (1977) Methods Enzymol. 47: 407-430) (mp 269-269.5° C (dec.) [lit. m.p.

272-273.5°C]) and toluene-3,5-dicarboxylic acid (Fittig et al. (1868)Anal. Chem. 147: 292-312) (mp 294.5-2960 C (water) [lit. m.p. 287-288° were prepared according to literature methods. DMF was distilled under N 2 from CaH 2 and stored over molecular sieve under N 2 before use. CCl 4 was fractionally distilled before use. Sulfonatoethyl methanethiosulfonate (la) was purchased from Toronto Research Chemicals (2 Brisbane Rd., Toronto, ON, Canada). All other chemicals were used as received from Sigma-Aldrich or Baker. All flash chromatography was performed using silica gel (Whatman, 60A, 230- 400 Mesh). Melting points were determined using an Electrothermal IA9000 series digital melting point apparatus and are uncorrected. IR spectra were recorded on Bomem MB or Perkin-Elmer FTIR Spectrum 1000 spectrophotometers. 'H NMR and "C NMR spectra were recorded on a Varian Gemini 200 NMR spectrometer at 200 and 50.3 MHz respectively. ES-MS data were acquired using a PE SCIEX API III Biomolecular mass spectrometer. All other MS and HRMS data, were acquired using Micromass 70-250S or Micromass ZAB-SE mass spectrometers according to the ionization methods indicated.

-38- )0 -W000/37658 PCTIUS99/30362 Microanalyses were performed by Canadian Microanalytical Service Ltd. (Delta, V4G 107, Canada). 'Solvents were removed .in vacuo.

7t 4-Carboxvbutyl methanethlosulfonate 0lb).

A solution of 5-bromopentanoic acid 238 g, 6. 84 mmol) and NaSSO 2

CH

3 916 g, 6. 84 mniol) in DMF (6.mL) was heated at 70*.C under N 2 After 2 h the solution was cooled, water (15 mL) added and the resulting mixture extracted with ether (30 nil x 3).

The organic fractions were combined, washed with brine, dried (MgSO 4 filtered and the.

solvent removed. The residue was purified by flash chromatography (ether: CH 2

CI

2 AeGH, 40:120: 1) to give lb 167 g, 80 as a white solid; mp 6 1-62. 50 C [lit. (U.S.

Patent 4,879,249), mp 69-71' IR (KBr) 1703 cm-1 1311, 1125 cm'1 (S-SO 2

'H

INMR (CDCI 3 8 1. 70-2. 00 (in, 4H, H-2, 2. 43 J 6. 9 Hz, 2H, 3. 20 J 6. 8 Hz, 2H, H-I1), 3. 34 3H, SSO 2 CHA) 8. 82 (br s, I1H, COON); "C NMR (CDCI 3 8 23. 4, 28. 9, 33. 1, 35. 9 ((CH 2 4 50. 7 (SSO 2 CHA) 178. 7 (COON); MS m/z 213 (M+HW, *1 195 2 O, 11), 133 115 (M+-CH 3

SO

2

-H

2 O, 100%); HRMS m/z Found 213. 0251 C 6 H1 3 0 4

S

2 requires 213. 0255.

methanethiosulfonate 1, I'-Carbonyidiimidazole 67 g, 0. 0411 rnol) was added to a solution of acid 364 g, 0. 0187 mol) in DMF (30 mL) and the resulting mixture stirred at 400 C under N 2 After 1. 5 h DBU 15 mL, 0. 041 mol) and I-BuOH (7.

7 rnL, 0. 0822 mol) were added. After 24 h the solution was cooled, ether (150 niL) added and the mixture acidified (HCI (aq. 1. 5 The ethereal layer was separated and the aqueous layer further extracted (ether, 150 mL). The organic fractions were combined, washed with water and 10% K 2 C0 3 (aq. ),dried (MgSO 4 filtered and the solvent removed.

The residue was purified by flash chromatography (EtOAc hexane, 1:50) to afford a colorless oil which solidified upon standing to give di-tert-butyl 58 g, 84%) as a white solid; mp 86. 5-87. 50 C (hexane); IR (film) 1712 cm" 1 1606, 1476 cm-' (Ar 'H NMR (CDC1 3 8 1. 60 18H, C(CH 3 2. 43 311, CH 3 7. 95 .(br s, 2H1, H-2, 8. 38 (br s, 1H, 1 3 C NMR (CDCI 3 8 21. 4 (CH 3 28. 2 (CLCH3) 3 81. 4 (C(CH 3 127. 7, 132. 1, 133. 7, 138. 1 165. 2 (COO).

NBS 521 g, 2. 93 mmol) and 1,1 '-azobis(cyclohexanecarbonitrile) (30 mg, 0. 12 mmol) were added to solution of this diester 712 g, 2. 44 mmol) in CC1 4 (10 niL) -39- WO 00(37658 ()7)PCr/US930362 and heated under reflux under N 2 After 3 h a second portion of initiator (30 Mgt 0. 12 rumol) was added. After a fur-ther 3 h the reaction solution was cooled and filtered. The.

filtrate was washed with sat. NaHCO3 (aq. dried (Mg9SO 4 filtered and the solvent O removed. The residue was partially purified by flash chromatography (EtOAc :hexane, 1:50) to afford crude 3,5-di(tert-butoxycarbo)benzylbromide 872 g, A solution 00 of 3 872 g, 2. 35 minol) and NaSSO 2

CH

3 327 g, 2. 44 mmol) in DMF (1 ruL) was heated at 500 C under N 2 After I h the reaction solution was cooled, diluted with water mL) and extracted with ether (15 ruL x The combined extracts were washed with brine, dried (MgSO 4 and the solvent removed. The residue was purified by flash chromatography (EtOAc :hexane, 1:8) to give 3,5-di(tert-butoxycarbo)benzyl methanethiosulfonate (0.

c-i570 g, 60%) as a colorless oil; IR (film) 1717 cm t 1604, 1477, 1456 cm 1 (Ar C=C), 1328, 1135 cm- 1

(S-SO

2 'H NMR (CDC1 3 8 1. 59 18H, C(CH 3 3 3. 07 3H,

SO

2

CH

3 43 2H, CH 2 8. 13 211, H-2, 8. 48 111, 'CNMR (CDCI 3 28. 2 (CLCH 3 3 40. 0 (CH 2 51. 3 (SO 2

CH

3 82. 1 (C-(CH3ij), 130. 2, 133. 2, 133.5, 135. 7 164.3 (COO).

A solution of 941 g,2. 30 mmnol) in CF 3 COOH CH 2

CI

2 (1 1 V/V, mL) was stirred at room temperature for 3 h, during which time a white precipitate was formed. The solvents were removed and the residue triturated with CH 2

CI

2 (5 mL). The resulting mixture was filtered, and the residue washed with CH 2

CI

2 and dried under vacuum to give Ic 611 g, 91 from 4) as a white solid; mp 199. 5-200* C (dec. IR (KBr) 1716, 1693 c6f' 1605, 1461 cm' (Ar 1319, 1128 cm- 1

(S-SO

2 'H NMR (acetoned 6 5 3. 29 3H, SO 2 CHA) 4. 69 211, CHA) 8. 36 J 1. 4 Hz, 2H, H-2, 8. 61 J 1. 7 Hz, I H, H14); 1 3 C NMR (acetone-d6) 8 40. 0 (CH 2 51. 2 (SO 2

CH

3 130. 8, 132. 5, 135.

4 2, 13 8. 5 166. 3 (COOH); MS m/z 290 273 2 10

CH

3

SO

2 H, 100), 179 (Mt SSQ 2

CH

3 HRMS m/z (FAB3+): Found 290. 9987 Clo~HI106S2 requires 290. 9998.

3.3-Dicarboxybutyl methanethlosulfonate (1d).

Anhydrous K 2

CO

3 67 g, 12. 0 mniol) was added to a solution of methyl Meldrum's acid (Sb) (I g, 6. 33 rnol) in DMF (33 mL) under N 2 and stirred vigorously.

After I h the supernatant liquid was added dropwise to a solution' of 1 ,2-dibromoethane 9 mL, 22.2 mmol) in DMF (I11 mL) under N 2 After 89 h TLC (EtOAc :hexane, 1:3) 0 -WO 00/37658 PCTIUS99/30362 indicated conversion of starting material (Rf 0. 3) to a-major product (Rf 0. The reaction mixture was added to water (100 mL) and extracted with ether (100 raL x The organic fractions were combined, dried (MgSO 4 filtered and the solvent removed. The residue was purified by flash chromatography (EtOAc :hexane, 1:4) to give 5-(2'-bromoethyl)-2,2,5trimethyl- 1,3-dioxocyclohexa-4,6-dione (6b) 183 g, as a white solid; mp 84-85* C (ether hexane); IR (KBr) 1738, 1784 cm-' NMR (CDC1 3 6 1. 66 3H, CH 3 1.

.76, 1. 78 (s x 2, 3H x 2, C(CH 3 2 2. 61 J 8Hz, 2H, 3. 32 J 8Hz, 2H, 1 3

C

NMR (CDCI 3 6 25. 2, 26. 6, 29. 1, 30. 1, 42. 4 (CH 3

C(CH

3 2 C-I1', 49. 4 106.

0 (C-(CH4 3 2 169. 5 MS rn/z 249, 251 (M+-C4 3 206, 208 3 2 14), 162, 164 (Mt C(O)OCO(CH- 3 2 42), 69 (M -C(O)OCO(CH 3 2

-CH

2 Br, 100%).

14aSSO 2

CH

3 (776 mg+, 5. 80 mmol) was added to a solution of 6b 18 g, 4.

46 minol) in DMF (40 mL) under N 2 and the resulting solution warmed to 500 C. After 29 h the reaction solution was cooled and the solvent removed. The residue was purified by flash chromatography (EtOAc hexane, 3:7 to give 2 2 2 ',5'-trimethyl-,3-dioxocyclohexa-4,6dionyl)ethyl inethanethiosulfonate (7b) l0 g, as a cloudy oil; IR (film) 173 7, 1771 n* 1300, 1133 cm-1 (5-502); 'H NMR (CDC 3 5 1. 68 3H, CH 3 1. 78, 1. 79 (s x 2, 3H x 2, C(CH 3 2 2. 47-2. 55 (in, 2H1, 3. 08-3. 16 (mn, 2H1, H-i1), 3. 34 3H,

*SSO

2 CHj); 1 3 C NMR (CDCI 3 6 24. 9, 29. 3, 30. 0, 31. 9, 49. 2 (CH 3

C(CH

3 2 C-I1, C-2), 49. 2 51. 2 (SSO 2 CHA) 106. 1 LC(CH 3 169. 5 MS rn/z 281

(M'-CH

3 269 239 (M+-C 3

H

5 O 159 (100), 141 113 103 87 (78), 69 (M+-C(O)0CO(CH 3 )-HS0C 3 79%).

Dowex 50W(H 4 resin 53 g) was added to a suspension of 7b 08 g, 3.

rniol) in p-dioxan 5 inL) and distilled water (35 mL) and stirred at room temperature.

After 68 h the reaction mixture was filtered and the solvent removed. The resulting solid was recrystallized from water acetone /ethyl acetate to give id (738 mng, 79%) as a white solid; mp 109-11 P C; IR (KBr) 1706 cnf' 1317, 1133 cm-' (S-SO 2 'H NMR (D 2 0) 6 1. 43 3H, H4), 2. 25-2. 33 (in, 2H, 3. 16-3. 24 (in, 2H, 3. 45 3H, SS02CH 3 1 3 C NMR (13 2 0) 6 20. 3 (CA4),32. 1, 36. 4 50. 5 (SSO 2

CH

3 54. 3 176. 0 (COOH); MS m/lz 256 (M 132 (M+H 4

-CH

2

SSO

2

CH

3 40), 116 87 HRMS rnfz Found 254. 9996 C 7 H41 10 6

S

2 requires 254.

9997.

WO.00/37654)7 PCT1VS99/30362 3.3.4-Trfcarboxvbutvl methanethiosulfonate (le).

0 Anhydrous K 2 C0 3 2 g, 8. 68 mmnol) was added to a solution of Meidrumi's acid (5a) (1 g, 6. 94 mmol) in DMF (20 niL) under N 2 and stirred vigorously. After 2 h the osupernatant liquid was added dropwise over a period of 1 h 30 min to a soluti 'on of tert- Z 5 butylbromoacetate 14 mL, 7. 63 nimol) in DMF (5 niL) under N 2 After a further 52 h TLC (acetone toluene, 1:9) indicated the conversion of starting material (Rf 0. 45) to major (RrO0. 5) and minor (Rf 0. 8) products. The reaction mixture was added to water (100 niL) and extracted with ether (100 mL x The organic fractions were combined, dried (MgSO 4 filtered and the solvent removed. The residue was purified -by flash chromatography (EtOAC hexane, 3:17 to 1:3) to give 5,5-di(tert-butoxycarbo)methyl-2,2dimethyl-1,3-dioxocyclohexa-4,6-dione (412 mg, nip .103-1050 C (ether hexane); 'H NMR (CDCl 3 8 1. 41 18H, C(CH 3 3 x 1. 92 6H, C(CH 3 2 2. 97 411, CHj 2 C00Bu' x 1 3 C NMR (CDC1 3 8 28. 5 (C(_QH 3 29. 2, 44. 1 (CH 2 CO0Bu', C(CH 3 2 47. 2 83. 1 LC(CH 3 3 108. 5 LC(CH 3 2 168. 0, 168. 9 (C4, C-6, CQOBu'); and a mixture of $a and 5-tr-uoyab~ehl22dmehl13dooylhx-,-in This mixture was purified by repeated crystallization from ether/hexane to give 5c (1.

g, 59%) as a white solid; mp 124-126* C (ether/ hexane); IR (KBr) 1772, 1755, 1712 cm- 'H NMR (CDC1 3 8 1. 43 9H, C(CH 3 3 1. 80 6H, C(CH 3 2 3. 09 J4Hz, 2H, Cii2C00Bu), 3. 70 J 4Hz, IlH, 13C NMR (CDC1 3 8 18. 7, 28. 8, 33. 0,43. 4

(QH

2 C00Bu', C(CH 3 2 28. 5 (CLCH 3 3 82. 8 (C(CH 3 3 105. 6 LC(CH 3 2 165. 6, 169. 6 C-6, COOBu'); MIS tn/z 257 100), 200 159 (25) 143 Anhydrous 1(2003 (300 mg, 2. 17 nimol) was added to a solution of 5c (400 mg, 1. 5 5 mmol) in DMF (10 niL) under N 2 and stirred vigorously. After I h the supernatant liquid was added dropwise to a solution of 1,2-dibromoethane 7 mL, 8. 06 nimol) in DMF (3 niL) under N 2 at 500 C. After 70 h, t. 1. c. (EtOAc hexane, 1:9) indicated the conversion of starting material (Rf 0. 1) to a major product (Rr 0. The reaction mixture was cooled, added to distilled water (50 niL) and extracted with ether (50 niL x The organic fractions were combined, dried (MgS0 4 filtered and the solvent removed. The residue was purified by flash chromatography (EtOAc hexane, 1:9) to give 5-(2 30 bromoethyl)-5-(tert-butoxycarbo)methyl-2,2-dimethyl- 1,3-dioxocyclohexa-4,6-dione (6c) (372 mg, 66%) as a white solid; mp 120-123' C (ether hexane); IR (KBr) 1773, 1731 cm7' 'H NM (ODC1 3 5 1. 40 9H, C(CH 3 3 80, 1. 93 (s x 2, 3H x 2, C(CH 3 2 2.

I2C -W 017658 PCT/US99/30362 41 J8Hz, 2H, H-iF), 3. 11 2H, Cii 2 C00Bu), 3.32 J8Hz, 2H, H-21); 1 3 C NNM (ODC1 3 8 25. 2, 29. 3,29. 9,41. 0,41.4 (CH 2 C00Bu', C(CH 3 2 28.

(C(-CH

3 3 51. 2 83. 2 (gCCH 3 3 107. 8 (C(0113)2), 167. 8, 170. 2 0-6, .!OQB ul); MS m/z 287 257 (M-(CH 2 2 Br, 100), 142 (15) 79, 81 (Be, 9 1%v).

NaSSO 2

CH

3 (143 mg, 1. 07 ramol) was added to a solution of 6c (301 mg, 0.

82 mmnol) in DMF (20 mL) under N 2 and the resulting solution warmed to 500 C. After 29 h the reaction solution was cooled and the solvent removed. The residue was purified by flash chromatography (EtOAc :hexane, 1:3) and crystallized from ether to give 2-(5'-(ter1butoxycarbo)methyl.2 '-dimethyl-1 3 -dioxocyclohexa.4,6..dionyl)ethyl methanethiosulfonate (7c) (280 mg, 86%) as a colorless crystalline solid; mp, 103-105o C (ether/hexane); IR (KBr) 1772, 1738, 1717 cm" 1314, 1129 cm"' (S-SO 2 'H1NMR (CDC1 3 8 1. 41 9H, 0(0C33)A1 83, 1. 93 (s x 2, 3H x 2, 0(0113)2), 2.33-2. 41 (in, 2H1, H1-2), 3. 10-3. 18 (in, 2H, H-I1), 3. 13 211, Cjj2CQOBu'), 3. 32 3H, 5502013) 3

C

NMR (ODCd 3 8 28. 0 (C(CH 3 3 28. 9,29. 2, 30. 7, 37. 9, 40. 2 (CH2COOBu',

C(CH

3 2

C-

1, 50. 0 50. 6 (SSO2CH 3 82. 8 (C(CH 3 3 107. 3 (C(CH 3 2 167. 2, 169. 7 (C- COOBa'); MS in/z 395 3 81 (M--CH 3 281 257 (M--(CH 2 2 SS0 2

CH

3 100), 215 158 A solution of 7c (138 mg, 0. 35 nmol) in CF 3 C00D: D 2 0 2 mL) was heated to 500 C. After 32 h, 'H NMR spectroscopy showed the conversion of starting material to a single product. The solution was cooled and the solvent removed. The residue was purified by flash chromatography (butan-1I-ol AcOH water, 4: 1: 1) and ion exchange chromatography (Amberlyst A21, 30% v/v CF 3 C00H (aq. as eluent) to give le (73 mg, as an amorphous solid; IR (KBr) 1706 cm"' 1310, 1127 cm-1 (S-SO 2 'H NMR (13 2 0) 8 2. 25-2. 34 (in, 2H, 3. 01 2H1, 3. 12-3. 20 (in, 2H, 3. 45 3H,

SSO

2

CH

3 3 C NMR (13 2 0) 8 34. 3, 37. 7, 41. 1 0-2, 52. 9 (55020113) 58. 0 (C- 177. 0, 177. 1 (COOH, CH2COOH); MS mnk 299 42), 221 183 111 91 Anal. calcd. for Cg1l2Og2: 0, 32. 00; H, 4. 03%; found: 0, 3 1.

84; H 3. 91 Site-specific Chemical Modification.

To approximately 25 mg of each of the SBL mutants in CHES buffer mL; 70 mM CHES, 5 mM MES, 2 MM CaC1 2 pH 9. 5) at 20' C was added each of the WO0037658 PCT/US99/30362 -i methanethiosulfonate reagents (100 tL of a 0. 2 M solution: lb in CH 3 CN: HzO la,c,d,e in water), in a PEG(MW 10,000)-coated polypropylene test tube and mixed using an end-over-end rotator. The progress of modification was followed using specific activity O measurement, monitored spectrophotometrically (10 pL aliquots in 0. 1 M Tris-HCI buffer, Z 5 pH 8. 6, 0. 005 Tween 80, and 1% DMSO, with succinyl-AAPF-pNA (1 mg/mL) as 00 substrate at 250 C, 6410 8800 cm'l) (Bonneau et al. (1991) J. Am. Chem. Soc. 119: 1026-1030) on a Perkin-Elmer Lambda 2 spectrophotometer. The reaction was terminated Swhen the addition of a further 100 pL of methanethiosulfonate solution gave no further change in specific activity, typically after 2 to 3 h. The reaction solution was purified on a 10 disposable desalting column (Pharmacia Biotech PD-10, Sephadex G-25 M) pre-equilibrated C with MES buffer (5 mM MES, 2 mM CaCI 2 pH 6. The CMM was eluted with this buffer 5 mL), dialyzed against MES buffer (10 mM MES, I mM CaCI 2 pH 5. 8, 1 L x 3) O at 40 C and subsequently flash frozen and stored at -18* C. The free thiol content of all CMMs, was determined spectrophotometrically by titration with Ellman's reagent (Ellman et al. (1961) Biochem. Pharmacol. 7: 88-95) (6412 13600 M- 1 in phosphate buffer 0.

M, pH 8. 0. In all cases no free thiol was detected. Modified enzymes were analyzed by nondenaturing gradient gels at pH 4. 2, run towards the cathode, on the Pharmacia Phast-system and appeared as a single band. Each of the CMMs showed reduced mobility relative to wild-type. Prior to ES-MS analysis CMMs were purified by FPLC (BioRad, Biologic System) on a Source 15 RPC matrix (17-0727-20 from Pharmacia) with acetonitrile, 0. 01% TFA as the running buffer and eluted with 80% acetonitrile, 0. 01% TFA in a one step gradient. MS nm/z (ES-MS): N62C-a (Berglund et al. (1997) J. Am. Chem. Soc.

119: 5265-5266) calculated 26826, found 26828; S156C-a (DeSantis et al. (1998) Biochem.

37: 5968-5973) calculated 26853, found 26859; S166C-a calculated 26853, found 26851; L217C-a (Berglund et al. (1997) J. Am. Chem. Soc. 119:5265-5266) calculated 26827, found 26828; N62C-b calculated 26819, found 26820; S156C-b calculated 26846, found 26846; S166C-b calculated 26846, found 26846; L217C-b calculated 26820, fourid 26820; N62C-c calculated 26897, found 26896; S156C-c calculated 26924, found 26928; S 166C-c calculated 26924, found 26928; L217C-c calculated 26898, found 26904; N62C-d calculated 26863, found 26870; S 56C-d calculated 26890, found 26892; S166C-d calculated 26890, found 26894; L217C-d calculated 26864, found 26866; N62C-e calculated -44- 00 0 WO 00/37658 PCT/US99/30362 26907, found 26909; S 156C-e calculated 26934, found 26939; S 166C-e calculated 26934, c found 26939; L217C-e calculated 26908, found 26911.

Active Site Titrations.

The active enzyme concentration was determined as previously described (Hsia et al. (1996) J. Anal. Biochem. 242: 221-227) by monitoring fluoride ion release upon enzyme reaction with -toluenesulfonyl fluoride (PMSF) as measured by a fluoride ion sensitive electrode (Orion Research 96-09). The active enzyme concentration determined in this way was used to calculate kcat values for each CMM except in the case of S 66C-c for Iwhich total protein concentration as determined by absorbance at 280 nm (E280 23000 M" 10 (Gron et al. (1990) Eur. J. Biochem. 194 :897-901) was used.

Kinetic Measurements.

Michaelis-Menten constants were measured at 25(t 0. 2)0 C by curve fitting (GraFit' 3. 03) of the initial rate data determined at eight or nine concentrations 125 mM-4. 0 mM) of succinyl-AAPF-pNA substrate in 0. 1 M Tris-HCI buffer containing 0.005% Tween 80, 1% DMSO, pH8.6 410 8800 cm' 1 Bonneau et al. (1991)J. Am.

Chem. Soc. 119: 1026-1030).

Example 2 Altering the Specificity of Subtilisin Bacillus lentus Through the Introduction of Positive Charge at Single Amino Acid Sites The use ofmethanethiosulfonates as thiol-specific modifying reagents in the strategy of combined site-directed mutagenesis and chemical modification allows virtually j unlimited opportunities for creating new protein surface environments. As a consequence of our interest in electrostatic manipulation as a means of tailoring enzyme activity and specificity, we have adopted this approach for the controlled incorporation of multiple negative charges at single sites in the representative serine protease, subtilisin Bacillus lentus (SBL). We now describe the use of this strategy to introduce multiple positive charges. A series of mono-, di- and triammonium methanethiosulfonates were synthesized and used to modify cysteine mutants of SBL at positions 62 in the S2 site, 156 and 166 in the S 1 site and 217 in the SI' site. Kinetic parameters for these chemically modified mutants (CMM) enzymes were determined at pH 8.6. The presence of up to three positive charges into the WO./37658 PCT/US99/30362

S

1 S' and S 2 subsites of SBL resulted in up to 77-fold lowered activity, possibly due to 0interference with the histidinium ion formed in the transition state of the hydrolytic reactions I catalyzed.

Z Results And Discussion 00 Synthesis of Alkylammonium Methanethiosulfonates Ib-d Previous work (Berglund et al. (1997) J. Am. Chem. Soc. 119: 5265-5266; Kenyon and Bruice (1977) Methods Enzymol. 47: 407-430) has demonstrated that, of the S-methods available (Douglass and Farah (1959) J. Org. Chem. 24: 973-975; Levitt and Levitt S i (1972) J. Org. Chem. 37: 332-334; Weidner and Block (1972) J. Med. Chem. 15: 564-567; Palumbo and Caputo (1981) Synthesis 888-890; Block and Zhao (1992) J. Org. Chem. 57: 5815-5817; Billard et al. (1996) J. Org. Chem. 61: 7545-7550), direct nucleophilic displacement of a primary alkyl bromide by methanethiosulfonate ion provides the most efficient method for the preparation of alkyl methanethiosulfonates. This general method was therefore adopted as the basis for the preparation of all of Ib-d. The singly-charged trimethylammonium MTS 1b (Ginsberg (1962)J. Med. Pharm. Chem. 5: 1364-1367.) was prepared from 2-bromoethyltrimethylammonium bromide and NaSSO 2

CH

3 in 57% yield.

The doubly-charged MTS reagent Ic was synthesized from the readily available dichloroalcohol 10 (Figure 6 showing scheme Initial attempts at constructing an untethered dipositive MTS failed as consequence of the hindered nature of the neopentyl-like system and therefore the less-hindered tethered system Ic was adopted as a target.

(Although untethered 2,2-bis(aminomethyl)- -bromo-N,N'-di-tert-butoxycarbonyl-propane was successfully prepared, all attempts at introducing methanethiosulfonate ion under a range of conditions failed).

Nitrogen was introduced through the treatment of a solution of 1 with an excess ofNaN 3 at 1300 C to give diazide 2 0, which was formed in a high degree of purity and used without further purification. Br(CH 2 2 0H protected as its tert-butyldimethylsilyl (TBDMS) ether 3 0 (Branchaud (1983) J. Org. Chem. 48:3531-3538; Vader et al. (1989) Tetrahedron. 45: 2131-2142) was used to.introduce an ethyl tether to 2. NaH was used to deprotonate the free alcohol in 2 and the resulting alkoxide anion was alkylated with 3 to give tethered diazide 4 in 55% yield.

WO 00/37658 PCT/US99/30362 SDiazide 4 0 was hydrogenated in the presence of Pd-black in MeOH and the free amine groups thus formed were protected by treatment of diamine 5 0 with Boc20 to K give dicarbamate 6 (72% yield over 2 steps from With the amine groups now suitably Sprotected the silyl ether moiety of 6 was selectively deprotected using TBAF-to give alcohol 7 0 in 93% yield. Treatment of 7 with MsCI and then LiBr allowed the formation of the primary bromide 9 0 via the corresponding mesylate 8 0 in 93% yield. The primary bromide group in 9 was displaced with methanethiosulfonate to give protected diammonium MTS 10 0 in 87% yield. Gratifyingly, deprotection of the Boc groups in 10 using TFA in DCM allowed the successful preparation of target dipositive MTS reagent Ic in 67% yield (35% overall yield from dichloroalcohol.

For the synthesis of triply-charged MTS ld (scheme 4 shown in Figure 7), pentaerythritol 11 was chosen as a cheap and readily available starting material. In order to introduce three amine groups and one methanethiosulfonate it was necessary to differentiate three of the four alcohol groups found in 11. This was readily achieved through the treatment of 11 with concentrated HBr in the presence ofrefluxing strong acid (Overburger et al. (1975) Macromolecules. 8: 31-36), conditions which allowed the formation of j tribromide 12 (D'yachenko and Lukima, (1966) Izv. Acad. Nauk. SSSR :2237) in 49% yield.

As for the synthesis of Ic, NaN 3 was used to introduce the required amine groups in masked form. Compounds containing high levels of nitrogen, such as triazide 12 0 present potential explosion hazards (Dunn et al. (1990) J. Org. Chem.55: 6368-6373) and therefore preparations avoided isolation through immediate reduction of a solution of 13 using the Staudinger reaction (Staudinger and Meyer (1919) Helv. Chim. Acta. 2: 635-646; Martin et al. (1995) Org. Prep. Proc. Intl. 27: 117-120). The iminophosphorane product was i hydrolyzed upon acidic work-up to the corresponding triaminoalcohol which was isolated and purified through crystallization as its trihydrochloride salt 14 (Litherland and Mann (1938) J. Chem. Soc. 1588).

Functionalization of the remaining alcohol group in 14 required protection of the three introduced amino groups. This was achieved through the condensation of 14 with PhCHO in the presence of Et 3 N which allowed the rapid formation of the triazadamantane (Dunn et al. (1990) J. Org. Chem.55: 6368-6373) in an excellent 99% yield. The free alcohol group in 15 was converted to a methanethiosulfonate group in an analogous manner to that used for the synthesis of Ic. Therefore, base-catalyzed mesylation of 15 and subsequent treatment with LiBr and then NaSSO 2

CH

3 gave protected triamino MTS 16 0 in -47- WO00/3765 PCT/US99/30362 "t 39% yield over 3 steps. Treatment of an ethereal solution of 16 with dilute HCI allowed 0 cleavage of the triaazaadamantane protecting group and crystallization of the target triammonium MTS reagent as its trihydrochloride salt Id (69% yield, 7% overall yield from O 11).

00 5 Preparation of Chemically Modified Mutants (CMMs) MTS reagents la-d were used to modify the chosen SBL cysteine mutants, N62C, S 156C, S 166C and L217C under conditions described previously (Berglund et al.

(1997) J. Am. Chem. Soc. 119: 5265-5266; Stabile et al. (1996) Bioorg. Med. Chem. Lett.6: j2501-2512; DeSantis et al. (1998) Biochem. 37: 5968-5973). These reactions proceeded rapidly and quantitatively, as judged by the monitoring of changes in specific activity and by titration of free thiols with Ellman's reagent (Ellman et al. (1961) Biochem. Pharmacol. 7: 88-95), respectively. The structure of the charged CMMs was confirmed by ES-MS. Nonreducing native PAGE was used to determine the purity of all the enzymes, which appeared as single bands. Consistent with the introduction of positive charge, each of the CMMs showed increased mobility in the direction of the cathode relative to WT. The active enzyme concentration of the resulting CMM solutions was determined by active site titration with atoluenesulfonyl fluoride (PMSF) using a fluoride ion-sensitive electrode (Hsia et al. (1996) Anal. Biochem. 242: 221-227).

Kinetic Effects of Site Speciflc Modification The effects of modification upon SBL were assessed by the determination of kca, and KM for the hydrolysis of succinyl-AAPF-p-nitroanilide (Suc-AAPF-pNA) at pH of 8.6. The kinetic parameters of the 16 CMMs generated are compared with those of WT and unmodified mutants in Table 2 and Figure 8.

Table 2. Kinetic Parameters' for Modified Enzymes Entry Enzyme Pocket R Level of kc, (s 1 KM (mM) kca,/Ku Charge (s 1 .SBL-WT 153 4 0.73 0.05 209 2 N62C S 2 H 0 174 9 1.90 0.20 92 ±11 a 1 b 1 c 2 103 5 73 2 92 3 1.00 0.10 0.86 0.05 1.06 0.07 103 ±11 85 87 6 WO 00/37658 PCT/US99/30362 6 d 3 98 ±3 1.17 ±0.08 84 6 7 L217C SI' H 0 41 1 0.80±0.04 51± 3 8 a 1 38 1 0.64 0.06 59 6 9 b 1 43 ±1 0.69 0.03 62 3 c 2 8.0± 0.2 2.94 ±0.28 -2.7 0.3 11 d 3 23 3 2.90 ±0.16 7.8 1.2 12 S156C S 1 H 0 125±4 0.85 ±0.06 147 11 13 a 1 90±2 0.73 ±0.04 123 7 14 b 1 68 2 0.74 0.04 92 c 2 64 1 0.76 0.04 85 16 d 3 46 ±1 0.81 ±0.05 57 4 17 S166C H 0 42 ±1 0.50±0.05 84±9 18 a 1 50±1 0.68 ±0.04 74 19 b 1 33 2 1.42 0.13 23 2 c 2 55±2 1.27±0.10 43 4 21 d 3 9.3 ±0.2 1.16 ±0.05 8.0 ±0.4 Michaelis-Menten constants were measured at 250 C according to the initial rates method in 90.1 M Tris-HCI buffer at pH 8.6, 0.005% Tween 80, 1% DMSO, Suc-AAPF-pNA as the substrate.

At position 62, in the S 2 pocket, mutation to cysteine reduces kca,/Ku by 2fold. The tolerance of the S 2 pocket for multiple charge at a single site that was previously observed for negatively charged CMMs (Example is also observed here and kcIKus for N62C-a-d are near-identical to that of N62C. The underlying k,,s and KMs of N62C-a-d are also near-identical to each other, although 2-fold lower than N62C. The lowered ka,,s of N62C-a-d indicate that introduced positive charges destabilize the transition state of catalysis although this destabilization does not increase further with increasing charge.

At position 217, in the St' pocket, mutation to cysteine causes a 4-fold decrease in kca/Ku and demonstrates that mutation at this site is intrinsically more dramatic.

Subsequent modification with singly positively charged MTS reagents la,b alters kea/KM little and the kt/Kms of L217C-a,b are only 1.2-fold greater than L217C. In stark contrast, the introduction of two and three positive charges dramatically decreases kca/KM. In fact, the kca/Kus of L217C-c,d are 77- and 27-fold lower than WT, respectively. These decreases are the result of both decreased k,,s that are up to 26-fold lower than WT for L217-c and incresaed KM s that are up to 4-fold greater than WT for L217C-d.

At position 156, in the SI pocket, mutation to cysteine decreases kat/Ku 1.4fold. From S156C to S156C-a to S156C-d kK/KM decreases monotonically to 3.6-fold lower than WT as the level of positive charge increases. This gradual tailoring of the specificity of -49- WO00/765n PCTIUS99/30362 7t jSBL away from hydrophobic substrate Suc-AAPF-pNA even at surface-exposed position S156 is consistent with a parallel trend for multiply negatively charged CMMs seen in C IExample 1.

O At position 166, in the Si pocket, mutation of the internally-oriented side Z i5 chain to cysteine decreases kcalKuM 3-fold. Subsequent modification with singly positively 00 charged MTS reagents la,b decreases kcat/KM further. In spite of the identical level of positive charge introduced the ka/KM of S66C-b is a dramatic 3-fold lower than that of S 166C-a and 9-fold lower than WT. This difference is largely a result of decreased substrate binding and the KM of S 166C-(CH 2 2 NMe 3 is 2-fold greater than that of S 166C- 10 S(CH 2 2 NH3 This may be attributed to the added steric bulk of peralkylated side chain

-S(CH

2 2 NMe3 as compared with the unalkylated side chain -S(CH 2 2 NH3 Modification with doubly positively charged MTS reagent Ic partially restores kca,/KM to 0 only 4-fold lower than WT. This exception to the general decreases in activity with increased positive charge may be a consequence of the added flexibility of tethered side chain c. This may allow the orientation of this side chain out of the Si pocket and towards external solvent in a manner analogous to that suggested by molecular modeling analysis for other charged CMMs. In contrast, the kcai/Ku of S166C-d, which bears a triply positively charged side chain, is severely lowered to 26-fold lower than WT. In contrast to the trend observed at position 217, the general decreases in kca/Ks at position 166 are largely due to decreases in kca, alone. In fact, from S166C-b to -d the KM decreases monotonically to only 1.6-fold greater than WT.

Conclusions In summary, we have devised short and efficient synthetic routes to two novel multiply charged methanethiosulfonates, Ic and d. Such compounds, as well as being of .i interest in our approach to the controlled tailoring of enzyme activity, may prove useful in the study of ion channels. The use of MTS reagents in techniques such as the substitutedcysteine accessibility method (SCAM) (Akabas et al. (1992) Science 258: 307-310; Akabas et al. (1994) Neuron 13: 919-927; Akabas et al. (1994) J. Biol. Chem. 269: 14865-14868) has allowed aspects of membrane ion channel topology and conformation to be determined.

In particular the use of charged MTS reagents has given an invaluable insight into ion specificity (Cheung and Akabas (1997) J. Gen. Physiol. 109: 289-299) and mechanism of action (Stauffer and Karlin (1994) Biochem. 33: 6840-6849; Yang et al. (1996) Neuron 16: 00 WO00/37658 PCT/US99/30362 113-122; Holmgren et al. (1996) Neuropharmacol. 35: 797-804; Huynh et al. (1997) J. Gen.

C' Physiol. 110: 229-242; Rassendren et al. (1997) EMBOJ. 16: 3446-3454).

C( Using our established methodology, we were able to fully modify the cysteine O thiols of SBL mutants, N62C, S156C, S166C, and L217C, with these reagents. Without exception, mutation and modification at all four sites led to reduced catalytic efficiency in the hydrolysis of Suc-AAPF-pNA.

Experimental Mutants of Subtilisin Bacillus lentus (SBL) were generated, and WT and mutant enzymes purified as described previously (Stabile et al. (1996) Bioorg. Med. Chem.

Lett.6: 2501-2512; DeSantis et al. (1998) Biochem. 37: 5968-5973). NaSSO 2

CH

3 (Kenyon and Bruice (1977) Methods Enzymol. 47: 407-430) (mp 269-269.50 C (dec.) [lit. m.p.

272-273.5 0 and toluene-3,5-dicarboxylic acid (Fittig et al. (1868) Anal. Chem. 147: 292-312) (mp 294.5-296° C (water) [lit. m.p. 287-2880 were prepared according to literature methods. DMF was distilled under N 2 from CaH 2 and stored over molecular sieve under N 2 before use. 2-aminoethyl methanethiosulfonate hydrobromide (la) was purchased from Toronto Research Chemicals (2 Brisbane Rd., Toronto, ON, Canada). All other chemicals were used as received from Sigma-Aldrich or Baker. All flash chromatography was performed using silica gel (Whatman, 60A, 230-400 Mesh). Melting points were determined using an Electrothermal IA9000 series digital melting point apparatus and are uncorrected. IR spectra were recorded on Bomem MB or Perkin-Elmer FTIR Spectrum 1000 spectrophotometers. 'H NMR and "C NMR spectra were recorded on a Varian Gemini 200 or Unity 400 NMR spectrometers at the frequencies indicated. ES-MS data were acquired using a PE SCIEX API III Biomolecular mass spectrometer. All other MS and HRMS data, were acquired using Micromass 70-250S or Micromass ZAB-SE mass spectrometers according to the ionization methods indicated. Solvents were removed in vacuo.

4.1 2-(Trimethvlammonlum)ethyl methanethiosulfonate bromide (Ib).

A solution of 2-bromoethyltrimethylammonium bromide (1.25 g, 5.06 mmol) and NaSSO 2

CH

3 (0.75 g, 5.60 mmol) in MeOH (10 mL) was heated under N 2 under reflux.

After 50 h the resulting solution was cooled to -18° C. The white solid formed was filtered from the mixture and recrystallized from EtOH to give lb (796 mg, 57%) a white crystalline solid; mp 157.5-158.50 C (EtOH) [lit.( Ginsberg (1962) Med. Pharm. Chem. 5: 1364-1367), -51- WO 0057658 WO 0037638 PCT/US99/30362 7t 1600 C (EtOH)]; 'H NMR (D 2 0, 200 MHz) 8 3.14 9H, N(CH 3 3 3.54 3H, C14 3

SO

2 C) 3.57-3.64 (mn, 2H, 3.67-3.77 (in, 2H, H-2).

4.2 5.5-Bis(amnomethvl)-3-oxo-hexyI methanethiosulfonate dlidrochlorlde 005 NaN 3 (10.78 g, 166 mol) was added to a solution of 2,2-bis(chloroniethyl)- Ipropan-lI-ol (1.98 g, 12.6 minol) in DMF (88 niL) under N 2 and the resulting suspension warmed to 1300 C. After 6 h the solution was cooled, poured into water (600 mL) and extracted with ether (100 mL then 5 x 50 mL). The organic fractions were combined, dried (MgSO 4 filtered and the solvent removed to give 2,2-bis(azidomethyl)-propan- I -ol (3) (2.10 g, 98%) as a yellow oil; IR (film) 3380 2102 (N 3 crn'; 'H NMvR (CDC 3 400 MHz) 8 0.93 3H, H1-3), 1.76 (br s, III, OH), 3.32 4H, -CH 2 3.47 2H, C NMR (CDC1 3 100 MHz) 8 18.3 41.1 55.7 (-CH 2 66.2 HRMS m/z Found 171.0970 (M H 4

C

5 H, IN 6 0 requires 171.0994.

NaB (480 mg, 80% dispersion, 16 mimol) was added to a solution of 3 g, 12.4 mmol) and Bu 4 N1 (228 mg, 0.62 inmol) in THF (50 mL) under N 2 After 20 min 2bromo-1-O-tert-butydimethylsilylethanol (4.6 g, 19.2 nimol) was added dropwise. After 22h, the solvent was reduced and die residue partitioned between ether (150 niL) and water inL). The aqueous layer was reextracted with ether (50 niL x The organic fractions were combined, washed with brine (40 mnL), dried (MgSO 4 filtered and the solvent removed. The residue was purified by flash chromatography (EtOAc :hexane, 1: 19) to give I -O-tert-butyldimethylsilyl-3-oxo-hexan- 1 -ol (2.24 g, 55%) as a colorless oil; IR (film) 2100 (N 3 C11-1; 'H NMR (CDC1 3 400 MHz) 8 0.05 6H, Si(CH 3 2 0.88 9H, SiC(CH 3 0.94 3H, 3.24 2H, JHH', 11.9 Hz, -CHH'N- x 3.26 (s, 2H, 3.28 2H, JH,H, 12.1 Hz, -CHH-'N- x 3.47-3.50 (in, 2H, H1-2), 3.72-3.74 (in, 2H, 'CNMR (CDC1 3 100 MHz) 5 -5.3 (Si(CH 3 2 18.1 (SiC(GH 3 18.4 25.9

(CLCH

3 3 41.2 55.8 (-CH 2 62.6, 73.0, 73.7 C-2, MS m/z 351 (M Na 4 329 (M H 4 Pd-black (70 mg) was added to a solution of 5 (1 g, 3. 05 nimol) in MeOH (27 mL) under N 2 The resulting suspension was thoroughly degassed and H12 introduced. After 18 h the suspension was degassed, N 2 introduced, filtered through celite (MeQH as eluant) and the solvent removed to give crude 5,5-bis(aminomethyl)- I -O-tert-butyldimethylsilyl-3- -52- 0 -WO 00/37658 PCTIUS99/30362 ___oxo-hcxan-1I-oI (800 mg, 2. 90 mmol). This was dissolve in dioxan I M aq. NaOH 20 niL) and the resulting solution cooled to 0* C. Boc 2 O 4 g, 6. 41 mmol) was added and after 30 min. the resulting mixture warmed to RT. After 3 h the reaction mixture was acidified to pH 4 with I M KHS0 4 (aq. and partitioned between EtOAc (300 mL) and water (100 niL). The aqueous layer was further extracted with EtOAc (200 niL). The organic fractions were combined, washed with sat. NaHCO 3 (aq. 100 niL), brine (100 niL), dried (MgSO4), filtered and the solvent removed. The residue was purified by flash I chromatography (EtOAc :hexane, 1:9 to 1:3) to give 5,5-bis(aminomethyl)- I -0-tertbuydmtysllN -itr-uoyabnl3oohxn I -ol 04 g, 72% over 2 steps) as a colorless oil; JR (film) 3360 1701, (amide 1508 (amide 11) cm"; 'H NMR (CDCI 3 400 MHz) 8 02 6H, Si(CHi) 2 0. 75 3H, 0. 81 9H, SiC(CH 3 1. 34 1 8H, OC(CH 3 3 x2 2. 83 (dd, 2H, .IH,NH 5. 3 Hz, JHH' 14. 0 Hz, CHH'N- x 3. 00 (dd, 2H, JH,NH 7. 9 Hz, JuHw 14. 1 Hz, -CHH'N- x 3. 16 2H, H-4), 39 2H, J4. 9 Hz, 3. 65 2H, J4. 9Hz, 5. 30 (br t, 2H, J6. 8 Hz, NH x 2); 3 NMR (CDC 3 100 MHz) 8 4 (Si(CH3) 2 18. 2 (SiC(CH 3 18. 8 25. 7 (SiC(-CH 3 3 28. 3 (OC(CH 3 )3 x 40. 2 44. 4 (-CH 2 62. 3, 72. 8, 76. 3 (C-I1, C-2, 78. 7 (OC-(CH 3 3 x 156. 7 (-NH(CO)O- x MS m/z 351 (M Na+, 3), 329 (M A solution of TBAF in THE (I M, 3. 7 mL, 3. 7 runol) was added dropwise to 20 a solution of 7 04 g, 2. 18 mmol) in THF (17 mQL under N 2 After 3 h, the solvent was removed. The residue was dissolved in EtOAc (200 mL) and washed with water (100 nml, x The aqueous layer was reextracted with EtOAc (100 niL x The organic fractions were combined, dried (MgSOA) filtered and the solvent removed. The residue was purified by flash chromatography (EtOAc hexane, 1: 1) to give butoxycarbonyl-3-oxo-hexanlol (735 mg, 93%) as a colorless oil; IR (film) 3355 (OH, NH), 1700 (amide 1520 (amide 11) cm" 'H NMR (CDC 3 400 MHz) 5 0. 80 3H, H-6), 1. 39 18H, OC(CH 3 3 x 2),12.93 (dd, 2H, JH,NH 5.6 6Hz, .IH,H- 14. 0 Hz, -CHH'N- x 3.

0 1 (dd, 2H, JH',NII 8. 0 Hz, JH,H' 13. 9 Hz, -CHH'N- x 3. 14 2H, 3. 29. I1H, OH), 3. 47 2H, J 4. 5 Hz, 3. 65 2H, J4. 3 Hz, H-I1), 5. 3 0 (br s, 2H, NH x 1 3

C

NMR (CDCI 3 100 MHz) a 18. 8 28: 3 (OC(CH 3 3 x 40. 3 44. 0 (-CH 2 61. 3, 72. 7, 74. 7 (C-i1, C-2, C4), 79. 2 (O-C(CH 3 3 X 156. 7 (-NH(CO)O- x MS m/lz 385 (M Na+, 45), 363 (M -WO 00137658 PCTIUS99/30362 MsCI 24 mL, 3. 10 mmol) was added dropwise to a solution of 8 (735 mg, 2. 03 ninol) and Et 3 N 57 niL, 4. 09 nunol) in DOM (5 mi.) under N 2 at O* C. After 1 h the solution was warmed to RT. After a further 16 h the solution was diluted with DCM o (100 washed with sat. NaHCO 3 (aq. 30 water (30 brine (30 dried (MgS0 4 filtered and the solvent removed. The residue was purified by flash 00 chromatography (MeOH CHC1 3 I1: 25) to give butoxycarbonyl-1-O-methanesulfonyl-3-oxo-hexan-1-oI (880 mig, 99%) as a colorless oil; IR (film) 3360 1700 (amide 1520 (amide 11) 1362, 1173 (0-SO 2 'H NMR

(CDCI

3 400 MHz) 8SO. 83 3H, 1. 41 18H1, OC(CH 3 3 x 2. 94 (dd, 2H, JH.N.

5. 9 Hz, JH,H' 14. 2 Hz, -CHjH'N- x 3. 05 3H, CH 3

SO

2 3. 06 (dd, 2H, JH'NH 7. 7 Hz, JH,H' 13. 9 Hz, -CHWHN- x 3. 24 2H, 3. 65-3. 67 (in, 2H, 4. 33-4. 35 (in, 2H, H-l1), 5. 27 (br t, 2H, J 6. 4 Hz, NH x 1 3 c NMR (CDCI 3 100 MHz) 8 18. 8 28.

4 (OC(CH 3 3 x 37. 7 (CHjSO 2 40. 5 44. 1 (-CH 2 68. 6, 69. 2, 76. 3 (C-I1, C-2, C4), 79. 2 (O-C(CH 3 3 x 156. 8 (-NH(CO)O- x MS m/z (FAB3+): 441 (M LiBr (860 mg, 9. 89 mmnol) was added to a solution of 9 (24 .5 mg, 0. 62 minol) in acetone (25 mL) under N 2 and heated under reflux. After 8 h the reaction mixture was cooled and the solvent removed. The residue was partitioned between ether (150 mL) and water (50 niL). The aqueous layer was reextracted with ether (50 mL x The organic frioswrcobnddre(gO, filtered and the solvent removed to give bis(aminomethyl)-lI-bromo-NN'-di-tert-butoxycarbonyl-3-oxo-hexane (10) (769 mig, 93%) as a yellow oil which was used directly in the next step.

NaSS0 2

CH

3 (315 mg, 2. 35 mmol) was added to a solution of 10 (769 mg, 1.

81 mniol) in DMF (30 mL.) and the resulting solution warmed to 500 C under N 2 After 20 h TLC (EtOAc :.hexane, 1: 1) showed the conversion of starting material (Rf 0. 8) to a major product (Rf 0. 35). The solvent was removed and the residue purified by, flash chromatography to give (EtOAc :hexane, 1: 1) butoxycarbonyl-3-oxo-hexyl methanethiosulfonate (11) (720 mig, 87%) as a colorless oil; IR (film) 33 85 1700 (amide 1508 (amide 11)1320, 1167 (S-SO 2 CM'1;'H NMR

(CDCI

3 400 MHz) 8 0. 83 3H, 1. 42 1 8H, OC(CH 3 3 x 2 2. 95 (dd, 2H, JH,N11 6. 1 Hz, .iw 14. 2 Hz, -CIIH'N- x 3. 05 (dd, 2H, JH',NHi 7. 2 Hz, JH,H' 14. 2 Hz, -CHHfNx 3. 24 211,1H4), 3.314 3H, CH 3

SO

2 3. 35 2H, J4. 9 Hz, 11-2), 3. 71 2H, J 4. 9Hz, 5.20 (brs, 211,NH x2); 'C NMR (CDC1 3 100 MHz) 8 18. 7 28.4 -54- WO 00/37658 PCT/US99/30362

(OC(CH

3 3 x 36. 2 40. 5 44. 2 (-CH 2 50. 6 (CH 3

SO

2 69. 6, 76. 2 (C-2, 79. 2 (OC(CH 3 3 x 156. 7 (-NH(CO)O- x MS m/z 479 (M Na 8), 457 (M H 11 (720 mg, 1. 58 mmol) was dissolved in DCM (20 mL) under N 2 and TFA (20 mL) was added. After 1 h the solvent was removed. The residue was purified by ion exchange chromatography (Dowex 50W(H), 4 x 3 cm, eluant aq. HCI, concave gradient 0.

5-2. 5 M) to give Ic as a white foam (348 mg, 'H NMR (D 2 0, 400 MHz) 5 0. 91 (s, 3H, 2. 89 2H, JH.H 13. 5 Hz, -CHH'N- x 2. 99 2H, JH.H' 13. 2 Hz, -CHH'Nx 3. 28 2H, J5. 6 Hz, 3. 31 3H, CH 3

SO

2 3. 42 2H, 3. 63 2H, 9Hz, "C NMR (D 2 0, 100 MHz) 17. 8 36.4, 37. 1, 45. 2 C-5, -CH 2 6 (CH 3

SO

2 70. 1, 75. 4 MS m/z 289 (M Na 4 15), 257 (M H HRMS m/z Found 257. 1000 (M C 8

H

21

N

2 0 3

S

2 requires 257. 0994.

4. 3 2 2 -Bis(aminomethvl)-3-aminopropvl methanethiosulfonate trihydrochloride (ld).

Pentaerythritol (12) (12. 8 g, 94 mmol) was dissolved in glacial HBr (1 5 v/v, 60 mL) and heated under reflux. After 24 h 40% HBr (50 mL) and c. H 2 S0 4 (23 mL) were added and the resulting solution heated under reflux. After a further 24 h the reaction mixture was cooled. The lower liquid layer from the resulting mixture was separated and dissolved in CHC 3 (50 mL), washed with water (20 mL), dried (anhyd. K 2

CO

3 filtered and the solvent removed. The residue was purified by flash chromatography (EtOAc hexane, 1:9) to give 2,2-bis(bromomethyl)-3-bromo-propan-l-ol (13) (15. 1 g, 49%) as a white crystalline solid; mp 67-690 C [lit., mp 68-69. 50 C (CCL4)J; 'H NMR (200 MHz, CDCI 3 8 1. 70 (br s, 1H, OH), 3. 55 6H, -CH 2 Br), 3. 75 2H, J6 Hz,

-CH

2 0H).

NaN 3 (15. 7 g, 241. 5 mmol) was added to a solution of 13 6 g, 20. 3 mmol) in DMF (120 mL) under N 2 and the resulting mixture warmed to 100* C. After 28 h the solution formed was cooled, poured into water (1 L) and extracted with Et2O (250 mL then 75 mL x The organic fractions were combined, dried (MgSO 4 filtered and the volume of solvent reduced to 100 mL. p-Dioxan (200 mL) was added and the volume of solvent reduced again to 100 mL. p-Dioxan (250 mL), PPh 3 (26. 6 g, 101. 4 mmol) and NH 3 30%, 100 mL) were added with stirring. After 19 h the solvent was removed, the WO*.0/3765 pCTUS99/303 6 2 residue suspended in CHC13 (400 niL) and extracted with HCI (aq. 5M, 75 niL x The aqueous fractions were combined, washed with CHC1 3 (20 niL x 4) and concentrated to a volume of 50 niL. c. HCI (aq. 10 mL) was added and the solution cooled to 40 C. The z white solid that crystallized from solution.was filtered, washed with cold c. HCI (aq. 3 00 5 mL), EtOH (3 niL), Et 2 O (20 ml x 5) and dried under vacuum to give 2,2bis(amiflometbyl).3amno-Propan-l-oI trihydrochloride (15) 81 g, 57%) as a white crystalline solid; nip 295-2980 C (dec. [lit., nip 298' C (dec. 'H NMR (200 MHz, D 2 0) 8 cs 3. 28 6H, -CH 2 3. 83 2H, -CH 2 Et 3 N 3 niL, 23. 6 nimol) was added to a suspension of 15 63 g, 6. 7 mmrol) in MeOH (13 niL) under N 2 PhCHO 1 niL, 20. 8 mmnol) was added to the resulting solution and heated under reflux. After 20 min the solution was cooled and the solvent removed. The residue was slurried with water (10 mL) and filtered to give 7- (hydroxymethyl)2,4,6triphenyl,3,5tiaaadamantane (16) 65 g, 99%) as a white solid; RrO. 5 (EtOAc: hexane, nip 91-930 C [lit., mp 92-95* 'H NMR (400 MHz, CDC1 3 8 2. 93 4H, -CH 2

-CH

2 2. 97 lH, OH), 3. 20 J 12. 9 Hz, 2H, -CH 2 3. J 13. 2 Hz, 2H, -CH 2 5. 42 2H, PhCH.), 5. 64 IH, PhCHeq), 7. 23-7. 83 (in, Ar).

MsCl 78 niL, 10. 1 mmol) was added dropwise to a solution of 16 65 g, 6. 63 mmol) and Et 3 N 9 niL, 13. 6 mmnol) in CH 2

CI

2 (15 miL) under N 2 at 00 C. After I h the reaction mixture was warmed to room temperature. After 21 h the resulting solution was diluted with CH 2

CI

2 (1 50 niL), washed with NaHCO 3 (aq. sat. 30 mL), water (30 mL) and brine (30 niL), dried (NgOW, filtered and the solvent removed. The residue was purified by flash chromatography (EtOAc hexane, 1:3) to give (2,4,6-triphenyl-1,3,5triazaadamantanyl)nlethyl methanesulfonate 44 g, 77%) as a white solid; Rr 0. 5 (EtOAc: hexane, nip 167.170' C (dec. IR (KBr) 1440 cm'1 1348, 1173 cm-f 1 (0-SO 2

'H

NMR (400 MHz, CDCI 3 8 2. 85 3H, CH 3 SO2-), 2. 96 2H, -CH 2 3. 30 J 12. 7 Hz, 2H, -CH2N-), 3. 50 2H, -CH2O-), 3. 54 J 13. 2 Hz, 2H, -CH 2 5. 45 2H, PhCHax,), S. 64 (s,1IH. PhCHcq), 7. 3 3-7. 80 (in, 15H, Ar); 1 3 C NMR (100 MHz, CDCl 3 8 26.

3 37. 1 (CH 3 SO2-), 45. 9, 54. 6, 7.3. 5 (-CH 2 -CH2N-) 75. 3, 82. 8 (PhCH-), 126. 126. 6, 127. 4, 127. 6, 128. 8, 129. 0,139. 1, 139. 4 HRMS m/z Found 476.

1996

C

27

H,

0

N

3 0 3 S requires 476. 2008.

0 WO 00/37658 pCTIUS99/303 62 LiBr 2 g, 36. 8 inmol) was added to a Solution of this niesylate 44 g, M1 10 imol) in dry acetone (30 mL) under N 2 and heated under reflux. After 22 h TLC (EtOAc hexane, 1:3) showed the loss of starting material (Rf 0. 5) and the formation of a major product (Rr 0. The reaction mixture was cooled and the solvent removed:. The residue was partitioned between Et2O (200 niL) and brine (50 niL), dried (MgSO4), filtered and the solvent removed. The residue was purified by flash chromatography (EtOAc hexane, 2:23) to give 7 -(bromomethyl)-2,4,6-triphenyl-1,3,5-triazaadamantane 82 mg, 78%) as a yellow oil; Rf 0. 8 (EtOAc :hexane, 1:3) which was used directly in the next step.

NaSSO 2 CH3 (440 mg, 3. 28 mmnol) was added to a solution of this bromide 1 g, 2. 38 mmol) in DMF (35 mL) under N 2 and the resulting solution heated to 800 C.

After 96 h TLC (EtOAc hexane, 1:3) showed the loss of starting material (RrO0. 8) and the formation of a major product (Rf 0. The solution was cooled and the solvent removed.

The residue was purifled by flash chromatography (EtOAc hexane, 1: 3) to give (2,4,6triphenyl- 1,3 tizadmnanlmt methanethiosulfoflate (17) (759 mg, 65%) as a colorless oil; IR (film) 1451 cm-' 1323, 1132 cm 1

(S-SO

2 'H NMR (400 MHz, CDC1 3 5 2. 63 2H, -CH2S-), 2. 98 2H1, -CH 2 3. 18 3H, CH 3

SO

2 3. 24 2H, J 13. 2HI-z, -CH 2 3. 51 2H, J 13. 2 Hz, -CH 2 5. 44 2H, PhCH.), 5. 63 I H, PhCH,), 7. 22-7. 83 (in, 15H, Ar); 1 3 C NMR (100 MHz, CDCl 3 8 26. 4 43. 0, 48. 7, 56. 6 (-CH 2

-CH

2 50. 3 (CH 3 SO2-), 75. 1, 82. 7 (PhQH-), 126. 5, 126. 6, 127. 5, 127.

6, 128. 9, 129. 0, 139. 1, 139. 3 HRMS m/z Found 492. 1768 (M+HW);

C

27

H

3 oN 3 0 2 S2 requires 492. 1779.

Conc. HC1 (aq. ,50 drops) was added to a solution of 17 (759 ing, 1. 54 Mnmot) in EtOHIEt 2 O (5/1 v/v, 24 mL) and the resulting solution cooled to 40 C. After 1 h, the white solid that crystallized from solution was filtered, washed with cold EtOH and Et 2

O

and dried under vacuum to give Id (358 mng, 69%) as a fine white powder; mnp 199-204'

C

(dec. IR (KBr) 2890, 1605, 1513 cm' (NH 3 1284, 1123 cm- 1

(S-SO

2 'H NMR (400 MHz, D 2 0) 8 3. 17 6H, -CH 2 3. 38 3H, CH 3 SO2-), 3. 44 2H, -CH 2 3 C NMR (100 MHz, CDCI 3 8537. 0 40. 7, 40. 8 (-CH 2

-CH

2 50. 6 (CH 3

SO

2

HRMS

m/z Found 228. 0844

C

6 H 18

N

3 0 2 S2 requires 228. 0840.

We 0037658 pCT/US99/30362 W 00/37658 0 Site-specific Chemical Modification.

To approximately 25 mg of each of the SBL mutants in CHES buffer >mL; 70 mM CHES, 5 mM MES, 2 mM CaC12, pH 9. 5) at 200 C was added each of the Z methanethiosulfonate reagents (100 pL of a 0. 2 M solution in water), in a PEG(MW 0 5 10,000)-coated polypropylene test tube and mixed using an end-over-end rotator. The progress of modification was followed using specific activity measurement, monitored spectrophotometrically (10 pL aliquots in 0. 1 M Tris-HCl buffer, pH 8. 6, 0. 005 Tween S 80, and 1% DMSO, with succinyl-AAPF-pNA (1 mg/mL) as substrate at 25* C, 410o= 8800 M' cm (Bonneau et al. (1991) J. Am. Chem. Soc. 119: 1026-1030) on a Perkin-Elmer (1 10 Lambda 2 spectrophotometer. The reaction was terminated when the addition of a further 100 uL of methanethiosulfonate solution gave no further change in specific activity, S Itypically after 2 to 3 h. The reaction solution was purified on a disposable desalting column (Pharmacia Biotech PD-10, Sephadex G-25 M) pre-equilibrated with MES buffer (5 mM MES, 2 mM CaC12, pH 6. The CMM was eluted with this buffer 5 mL), dialyzed against MES buffer (10 mM MES, 1 mM CaCI2 pH 5. 8, 1 L x 3) at 40 C and subsequently flash frozen and stored at -180 C. The free thiol content of all CMMs, was determined spectrophotometrically by titration with Ellman's reagent (6412 13600 M-1 cm in phosphate buffer 0. 25 M, pH 8. 0. In all cases no free thiol was detected. Modified enzymes were analyzed by nondenaturing gradient gels at pH 4. 2, run towards the cathode, on the Pharmacia Phast-system and appeared as a single band. Each of the CMMs showed increased mobility relative to wild-type. Prior to ES-MS analysis CMMs were purified by FPLC (BioRad, Biologic System) on a Source 15 RPC matrix (17-0727-20 from Pharmacia) with 5% acetonitrile, 0. 01% TFA as the running buffer and eluted with acetonitrile, 0.01% TFA in a one step gradient. MS m/z (ES-MS): N62C-a (Berglund et al.

(1997) J. Am. Chen. Soc. 119: 5265-5266) calculated 26763, found 26764; S156C-a (DeSantis et al. (1998) Biochem. 37: 5968-5973) calculated 26790, found 26791; S166C-a calculated 26790, found 26784; L217C-a (Berglund et al. (1997) J. Am. Chem. Soc.

119: 5265-5266) calculated 26764, found 26764; N62C-b calculated 26805, found 26808; S156C-b calculated 26832, found 26835; S166C-b calculated 26832, found 26835; L217C-b calculated 26806, found 26808; N62C-c calculated 26863, found 26863; S156C-c calculated 26890, found 26892; S166C-c calculated 26890, found 26899; L217C-c calculated 26864,

C)

00 WO 00/37658 PCT/US99/3036 2 found 26869; N62C-d calculated 26834, found 26835; S156C-d calculated 26861, found c 26866; S166C-d calculated 26861, found 26862; L217C-d calculated 26835, found 26837.

Active Site Titrations.

SThe active enzyme concentration was determined as previously described (Hsia et al. (1996) Anal. Biochem. 242: 221-227) by monitoring fluoride release upon enzyme reaction with at-toluenesulfonyl fluoride (PMSF) as measured by a fluoride ion sensitive electrode (Orion Research 96-09). The active enzyme concentration determined in this way was used to calculate values for each CMM.

Kinetic Measurements.

Michaelis-Menten constants were measured at 25(± 0. C by curve fitting (GraFit® 3. 03) of the initial rate data determined at eight or nine concentrations 125 mM-4. 0 mM) of succinyl-AAPF-pNA substrate in 0. 1 M Tris-HCl buffer containing 0.

005% Tween 80, 1% DMSO, pH 8. 6 (410 8800 (Bonneau et al. (1991) J. Am.

Chem. Soc. 119: 1026-1030).

Example 3 Tailoring The Specificity of the St Pocket of Subtilisin B. lentus: Chemical Modification Sof Mutant Enzymes as a Strategy for Removing Specificity Limitations In both protein chemistry studies and organic synthesis applications, it is desirable to have available a toolbox of inexpensive proteases with high selectivity and diverse substrate preferences. This example describes the generation of a series of chemically modified mutant enzymes (CMMs) of subtilisin B. lentus (SBL) possessing 1 expanded Si pocket specificity. Wild-type SBL shows a marked preference for substrates with large hydrophobic P 1 residues, such as the large Phe PI residue of the standard suc- AAPF-pNA substrate. In order to confer more universal PI specificity on SI, a strategy of chemical modification of mutant enzymes was applied. For example, WT-SBL does not readily accept small uncharged PI residues such as the -CH 3 side chain of alanine.

Accordingly, with a view to creating a S pocket that would be of reduced volume providing a better fit for small PI side chains, a large cyclohexyl group was introduced by the CMM approach at position S166C with the aim of partially filling up the S pocket. The S166C-S-

CH

2 -c-C 6 HIi CMM thus created showed a 2-fold improvement in kcat/K with the sucpCT1US9913 0362 WO 00/37658 S AAPA-pNA substrate and a 51-fold improvement in suc-AAPA-pNA/suc-AAPF-pNA Sselectivity relative to WT-SBL. Furthermore, WT-SBL doesnot readily accept positively or negatively charged PI residues. Therefore, to improve SBL's specificity toward positively Sand negatively charged P, residues, we applied the CMM methodology to introduce complementary negatively and positively charged groups respectively at position S166C in SI. A series of mono-, di-, and tri-negatively charged CMMs were generated and all showed improved kcat/KMS with the positively charged

P

1 residue containing substrate, suc-AAPR- SpNA. Furthermore, virtually arithmetic improvements in ktI/KM were exhibited with increasing number of negative charges on the S166C-R side chain. These increases culminated in a 9-fold improvement in kctlKM for the suc-AAPR-pNA substrate and a 61fold improvement in suc-AAPR-pNA/suc-AAPF-pNA selectivity compared to WT-SBL for the tri-negatively charged S 166C-S-CH2 CH2C(COO-)3 CMM. Conversely, the positively charged S166C-S-CH 2

CH

2 NH3 CMM generated showed a 19-fold improvement in k tKM for the suc-AAPE-pNA substrate and a 54-fold improvement in suc-AAPE-pNA/suc-AAPFpNA selectivity relative to WT-SBL.

In this example, we exploit strategy of applying a combination of site-directed mutagenesis and chemical modification to modify enzyme specificity. This approach This approach is illustrated in Scheme 5 (Figure 9) and entails the introduction of a unique cysteine residue at a selected position, followed by its chemical modification with methanethiosulfonate (Kenyon and Bruice (1977) Meth. Enzymol. 47: 407-430; Wynn and Richards (1995) Meth. Enzymol. 251: 351-356; Brocklehurs (1979) Int. J. Biochem. 10: 259- 274) reagents (MTS, la-lj) to generate chemically modified mutant enzymes (CMMs). The combination of site-directed mutagenesis and chemical modification has previously been recognized as a powerful tool for the creation of new active-site environments (Sorensen et al. (1993) Biochemistry 32: 8994-8999; Bech and Breddam (1988) Carlsberg Res. Commun.

53: 381-393), in mechanistic studies (Gloss and Kirsch (1995) Biochemistry 34: 12323- 12332; Smith and Hartman (1966) J. Biol. Chem. 263(10): 4921-4925), for the investigation of protein packing (Wynn et al. (1996) Protein Sci. 5: 1026-1031), and for cofactor incorporation (Kuang et al. (1996) J. Am. Chem. Soc. 118: 10702-10706). This approach has also been applied to detailed studies of ion-channel properties (Foong et al. (1996) Biochemistry 36: 1343-1348; Holmgren et al. (1996) Neuropharmacology 35: 797-804; Yang et al. (1996) Neuron 16: 113-122) for site-directed introduction of spin-labels (Hubbell WO00/37658 PCT/US99/30362 et al. (1996) Structure 4: 779-783; Lin et al. (1998) Science 279(5358): 1925-1929), to probe CM receptor binding (Heinonen et al. (1998) Bioconjugate Chem. 9: 358-364), and in j investigations of membrane spanning proteins (Akabas et al. (1994) Neuron 13: 913-927; SChen et al. (197) Biochemistry 36: 1479-1486).

The subtilisin from Bacillus lentus (SBL, EC 3.4.21.14) is well suited as an exploratory vehicle for evaluating the potential of this combined site-directed mutagenesis chemical modification approach since it is a well characterized enzyme and is of synthetic (Shao and Arnold (1996) Cur. Opin. Struct. Biol. 6: 513-518; Lloyd et al (1998) i Tetrahedron: Asymmetry 551-561) as well as industrial (van der Osten et al. (1993) 10 Biotechnol. 28(1): 55-68) interest. Furthermore, SBL's high resolution crystal structure has been solved (Knapp et al. Brookhaven Database Entry 1JEA; Kuhn et al. (1998) Biochemistry 37 13446-13452), it has been cloned, over expressed and purified (Stabile et al. (1996) Bioorg. Med. Chem. Lett. 6(21): 2501-2506), and its kinetic behavior well characterized (Gron: H. et al. (1992) Biochemistry 31(26): 6011-6018; Egmond et al.

(1994) Protein Eng. 793-800; Maurer et al. (1996) Adv. Exp. Med. Biol. 379: 243-256; Olsen et al. (1996) Adv. Exp. Med. Biol. 379: 235-241). In addition, and importantly, wild type (WT) SBL contains no natural cysteine residues, and methanethiosulfonate reagents therefore react only with the introduced cysteine residue. The validity of the CMM approach for altering the stability (Gron et al. (1990) Eur. J. Biochem. 194: 897-901), specificity (Bech et al. (1993) Biochemistry 32: 2845-2852), kinetic properties (Berglund et al. (1996) Bioorg.

Med. Chem. Lett. 6(21): 2507-2512; Berglund et al. (1997) J. Am. Chem. Soc. 119: 5265- 5266; DeSantis et al. (1998) Biochemistry 37: 5968-5973; Plettner et al. (1998) Biorg. Med.

Chem. Lett. 8(17): 2291-2296), and pH-profiles (DeSantis and Jones (1998) J. Am. Chem.

?i Soc. 120(34): 8582-8586) of subtilisins has been recognized.

WT-SBL has a marked preference for substrates with large hydrophobic uncharged P, residues. In this study, we explore tailoring of the S, pocket of SBL to also accept small hydrophobic, positively charged, and negatively charged PI residues. In order to achieve this broadened PI tolerance, a simplistic strategy of steric and electrostatic complementarity was applied (Fersht (1985) A. Enzyme Structure and Mechanism: 2nd ed.; W.H. Freeman and Company: New York). Employing the crystal structure of SBL as our guide (Knapp et al. Brookhaven Database Entry 1JEA), the Serl66 residue, which is located at the bottom of the St pocket and whose side chain points inward toward the pocket, was chosen for mutagenesis to cysteine and subsequent chemical modification. Firstly, to expand -61- WO'00/37658 PCTJS99/30362 SBL's specificity toward small uncharged Pi residues, such as the small P 1 Ala residue of the Ssuc-AAPA-pNA substrate, we introduced large moieties at position 166 in SI, such as benzyl decyl cyclohexyl and steroidyl groups with a view to reducing the volume 0 of Si and inducing a better fit of small PI groups, thereby conferring elastase-like (Bode et al. (1989) Biochemistry 28(5): 1951-1963) substrate specificity on SBL. Then, to expand SSBL's specificity toward positively charged PI residues, such as the PI Arg residue of the suc-AAPR-pNA substrate, we introduced negatively charged groups at position S 166C in Si, c such as the ethylsulfonato moiety, and the dicarboxylic aromatic and aliphatic Smono- di- and tri aliphatic groups, to elicit complementary electrostatic attractions with a view to making SBL trypsin-like in its specificity (Perona et al. (1995) Biochemistry S7(34): 1489-1499). Conversely, to expand SBL's specificity toward negatively charged Pi residues, such as the negatively charged PI Glu residue of the suc-AAPE-pNA substrate, we 0 introduced the positively charged ethylamino group at position S 166C in Si.

Results: The preparations of the requisite MTS reagents Ic (Berglund et al. (1997) J.

Am. Chem. Soc. 119: 5265-5266), le-f and Id (Berglund et al. (1997) J. Am. Chem.

Soc. 119: 5265-5266; Examples 1 and li-j are as reported, and the steroidyl MTS i reagent Ig was prepared from cholic acid by the same methodology.

Each of the CMMs obtained was characterized in order to establish its purity and integrity. Titration of the CMMs with Ellman's reagent showed a residual thiol content of less than 2% in all cases, demonstrating that the MTS reactions were virtually quantitative. Mass analyses of the CMMs by electrospray mass spectrometry were consistent (1 6 Da) with the calculated masses. The purities of the modified enzymes were assessed by native-PAGE and in all cases only one band was visible. Furthermore, as expected relative to WT, the negatively charged CMMs SI66C-S-b,-d, and -i to -j displayed retarded mobility in the direction of the cathode, while the positively charged S166C-S-a CMM displayed greater mobility. That modification of cysteine is wholly responsible for altered activity was established by the absence of reaction of WT-SBL with the MTS reagents. Also, the modifications are fully reversible by treatment of each of the CMMs with p-mercaptoethanol, further verifying that chemical modification at cysteine was solely responsible for the observed changes in activity. The total amount of active enzyme was

G

MBiochem. 242: 221-227).

(N

Initially, three CMMs SI 66C-S-a, -b and with a positive, a negative, and with a large hydrophobic side chain respectively, were subjected to a kcat/KM screen with each of the test substrates, suc-AAP-F/A/R/E-pNA in order to identify any induced complementary electrostatic or improved hydrophobic interictions (Figure 10). While as expected k.t/KMs with the standard suc-AAPF-pNA were lowered, the kt/Ks of the CMMs whose Si sites were tailored toward the Ala, Arg, and Glu P 1 residues, improved with the appropriate substrate. This is illustrated in Figure 10 in the higher activity of SI 66C-S-c with suc-AAPA-pNA, of S166C-S-b with suc-AAPR-pNA, and ofS166C-S-a with suc- AAPE-pNA, all relative to WT.

Following the validation of the general design strategy from this initial screen, more complete kinetic analyses were undertaken. The substrate specificity of each of the SCMMs was evaluated kinetically with the standard large hydrophobic P residue containing substrate, suc-AAPF-pNA. In addition, the S166C CMMs modified with the large hydrophobic MTS reagents Ic, e-g were evaluated with the small hydrophobic P 1 residue containing substrate, suc-AAPA-pNA. The S166C CMMs modified with the negatively charged MTS reagents Ib, 1d, sh-lJ were evaluated with the positively charged P residue containing substrate, suc-AAPR-pNA. The S 1 66C CMM modified with the positively charged MTS reagent la was evaluated with the negatively charged P, residue containing substrate, suc-AAPE-pNA. The results are summarized in Table 3.

Table 1: Kinetic Evaluation(a) of Altered S, Pocket Specificity Entry Enzyme Substrate K mM k s wh k,/Kga mM' c WT suc-AAPF-pNA 0.73 0.08 153 k 4 209 2 S166C-S-a suc-AAPF-pNA 0.68 0.04 50 in 74 3 SI66C-S-b suc-AAPF-pNA 1.34 0.08 25.0 0.7 19 1 1 4 S166C-S-c suc-AAPF-pNA 1.17 0.06 23.1 ±0.5 20 1 S166C-S-d suc-AAPF-pNA 1.6 0.2 47 3 29 4 6 S166C-S-e suc-AAPF-pNA. 1.09 0.07 82 2 75 7 S166C-S-f suc-AAPF-pNA 0.70 1 0.05 4.8 0.1 6.90 0.05 8 S166C-S-g suc-AAPF-pNA 0.74 ±0.07 29 1 41 4 9 SI66C-S-h" suc-AAPF-pNA 1.52 0.06 48 1 31 1 S166C-S-1 suc-AAPF-pNA 2.26 0.10 67 2 30 2 11. S166C-S-J(b suc-AAPF-pNA 2.46 ±0.11 76 2 31 2 12 WT suc-AAPA-pNA 2.0 ±0.1 17.7 0.3 8.8 0.4 -63- WO.00/37658 CTIUS99/30362 13 S166C-S-c suc-AAPA-pNA 0.8 ±0.1 6.8 10.3 9 1 1 14 S166C-S-e suc-AAPA-pNA 1.90 ±0.03 6.8 *0.4 3.6 0.6 S166C-S-f suc-AAPA-pNA 1.90+ 0.07 28.2 0.4 14.8 0.6 16 S166C-S-g suc-AAPA-pNA 1.74 0.04 9.65 0.07 5.54 0.3 17 WT suc-AAPR-pNA 7.2 0.7 0.16 1 0.01 0.022 4 0.002 18 S166C-S-b suc-AAPR-pNA 3.4 0.3 0.17 0.01 0.050 ±0.005 19 S166C-S-d suc-AAPR-pNA 5.5 1.1 0.68 0.08 0.12 10.03 00 20 S166C-S-h suc-AAPR-pNA 8.2 0.9 0.35 0.02 0.041 0.005 21 S166C-S-i suc-AAPR-pNA 5.3 4 0.5 0.43 0.02 0.080 0.008 22 S166C-S-j suc-AAPR-pNA 5.2 0.6 1.06 1 0.07 0.20 1 0.03 23 WT suc-AAPE-pNA 4.4 0.4 1.75 0.08 0.40 0.04 C24 S166C-S-a suc-AAPE-pNA 1.9 10.1 14.5 ±0.3 7.6 0.4 S(a) Michaelis-Menten constants were measured by the initial rates method in pH 8.6 Tris- HCI buffer at 250 C with suc-AAPF-pNA as the substrate.

Discussion.

C 5 The significant substrate preference of WT-SBL for large hydrophobic Pi residues is apparent from its preference for the Phe P, residue of the standard suc-AAPFpNA substrate, by a factor of 9500-fold over the small Pi residue of suc-AAPA-pNA, by a factor of 24-fold compared to the positively charged PI residue of suc-AAPR-pNA, and by a factor of 522-fold compared to the negatively charged PI residue of suc-AAPE-pNA (Table 3, entries 1,12,17 and 23). These kinetic differences are due to changes in both binding, as reflected by KM, and in turnover number, Moreover, and predictably, the WT enzyme is by far the best catalyst with suc-AAPF-pNA, and its conversions to any of the CMMs were deleterious with respect to this substrate and resulted in kca,/KM decreases of up to 34-fold (Table 3, entries 2-11).

To improve the substrate specificity of SBL toward small hydrophobic Pi residues such as Ala, the simplistic approach of filling up the S, binding cleft was addressed by preparing the S 66C-S-CH 2

C

6 H5 S166C-S-CH 2

(CH

2 )sCH 3 S166C-S-CH 2

C

6

H,

and S166C-S-steroidyl CMMs. This design strategy attempted to mimic the function of the bulky Si-pocket side chains of ao-lytic protease (Bone et al. (1989) Biochemistry 28: 7600-7609; Bone et al. (1991) Biochemistry 30: 10388-10398; Bauer et al.

(1981) Eur. J. Biochem. 120: 289-294), and of elastase (Bode et al. (1989) Biochemistry 28(5): 1951-1963), which are responsible for their substantial preference for the small Pigroup containing suc-AAPA-pNA substrate over the large PI-group containing suc-AAPFpNA substrate (Bone et al. (1989) Biochemistry 28: 7600-7609; Bone et al. (1991) 00 0 -WO 00/37658 PCT/US99/30362 S.Biochemistry 30: 10388-10398; Bauer et al. (1981) Eur. J. Biochem. 120: 289-294; Bode et Sal. (1989) Biochemistry 28(5): 1951-1963). These CMMs (S166C-S-c,-e,-f, Table 3, entries 12-16) were then evaluated with the suc-AAPA-pNA substrate. All revealed slightly improved binding compared to WT, with the greatest improvement in KM being 2-fold for the S 166C-S-CH 2

C

6

H

5 CMM. However, of these four CMMs, only S 166C-S-CH 2 CsHIz showed both an improved kct and an improved kt/KM. While this design strategy yielded only one CMM with an increased preference for the small Ala PI residue, all of these modifications effectively excluded the larger Phe PI residue preferred by WT-SBL (Table 3, entries 4, Overall the selectivities with respect to kcai/KM for the suc-AAPA-pNA .1 10 substrate compared to the suc-AAPF-pNA substrate were improved by I1-fold for S 166C-S-

CH

2

C

6 Hs 1. 1-fold for S166C-S-CH 2

(CH

2 )sCH 3 51-fold for S166C-S-CH 2

C

6

HI

and 3. 2-fold for S166C-S-steroidyl all compared to WT. These differences in P, Ala selectivity may be a reflection of the orientation of the R side-chain of the CMM, with the side-chains of S166C-S-CH2C 6

H

5 and -CH 2

C

6

HI

1 behaving as though directed into the pocket and favoring P| Ala, whereas the side chains of S 166C-S-CH 2

(CH

2 )sCH 3 and S166C-S-steroidyl behave as though directed outward, thus not significantly altering the shape of the S pocket.

The above improvements in PI Ala acceptance, although modest, are encouraging and demonstrate the effectiveness of the adopted strategy. Tailoring the steric complementarity between enzymes and substrates has already been found to be challenging and comparison of these CMM results with previous literature studies targeting the same goal are both interesting and intriguing. For example, the G 166I mutation of subtilisin BPN' effected an almost 1000-fold decrease in kcat/K with the P =Phe suc-AAPF-pNA substrate.

However, this same G166I mutant, which was by far the most effective mutation, elicited a 10-fold improved kct/KM compared to WT with the Pi=Ala suc-AAPA-pNA substrate (Estell et al. (1986) Science 233: 659-663). Notably, the G166I subtilisin mutant is more selective for Ala over Phe than is our most selective CMM. However, in both cases the increases in selectivity are due mainly to decrease in k,.C/KM for suc-AAPF-pNA rather than increases in kt /KM for suc-AAPA-pNA. Both the results for both CMM and SDM approaches agree that decreasing the selectivity of an enzyme for a large hydrophobic residue containing substrate can be accomplished in a relatively facile manner by the introduction of large amino acid in the enzyme pocket, but that increasing the selectivity of an enzyme for a small hydrophobic residue containing substrate is much more difficult.

WO 00/37658 PCTIUS99/0362 Similarly, the G127A mutant of subtilisin YaB, whose specificity was already elastase-like, O effected a 10-fold improvement in ktJKM with the suc-AAPA-pNA substrate (Mei et al.

(1998) Protein Eng. 11(2): 109-117). However, the G127V mutant of subtilisin E induced a decrease in kct/KM with the suc-AAPA-pNA substrate, identifying an inconsistency in the SDM strategy (Takagi et al. (1996) FEBSLett. 395: 127-132). Thus the CMM approach 00 offers a complementary alternative to conventional site-directed mutagenesis toward the goal of tailoring the steric complementarity between enzymes and substrates (Takagi et al. (1997) Protein Eng. 10(9): 985-989).

Improving the substrate specificity of SBL toward positively charged P, residues such as Arg, was based on mimicking the common motif in trypsin-like enzymes (Knapp et al. Brookhaven Database Entry 1JEA) of high negative charge density of acidic residues that favor binding of positively charged substrate structures (Perona et al. (1995) Biochemistry 7(34): 1489-1499; Nakayama (2997) Biochem J. 327: 625-635). This goal was K1 addressed by S 66C-S-CH 2

CH

2

SO

3 S 166C-S-CH 2

(CH

2 2

CH

2 COO- S166C-S-

CH

2

C

6

H

4 -3,5-(COO) 2 S 166C-S-CH 2

CH

2

C(CH

3

)(COO)

2 and S166C-S-

CH

2

CH

2

C(COO-)

3 a series of CMMs which provide a S 1 pocket which is potentially mono-, di-, and tri-negatively charged. Evaluation of each of these CMMs with the suc- AAPR-pNA substrate revealed KMS that were up to 2-fold improved compared to WT (Table 3, entries 17-22). The general success of this approach is evident since all of the CMMs with a negatively charged -R side-chain showed improved activity compared to WT with an up to 7-fold improved kta and an up to 9-fold improved kcl/KM with the suc-AAPR-pNA substrate (Table 1: entries 17-22). Overall, the selectivities, with respect to kcat/KM, for the suc-AAPRpNA substrate compared to the suc-AAPF-pNA substrate were improved 25-fold for S166C-

S-CH

2

CH

2

SO

3 13-fold for S166C-S-CH 2

(CH

2 2

CH

2 COO- 39-fold for S166C-S- CH2C6H4-3,5-(COO) 2 25-fold for S166C-S-CH 2

CH

2

C(CH

3

)(COO-)

2 and 61-fold for S 166C-S-CH 2

CH

2 C(COO-) relative to WT.

The strategy of introducing charge complementarity to induce trypsin-like P 1 specificity in subtilisins has previously been explored using site-directed mutagenesis (SDM) (Wells et al. (1987) Proc. Nat. Acad. Sci. USA, 84: 1219-1223; Ballinger et al. (196) Biochemistry 33: 13579-13585; Bonneau et al. (1991) J. Am. Chem. Soc. 113: 1026-1030).

Interestingly however, the G 166D and G 166E mutants of subtilisin BPN' caused decreases in k.t/KM with the suc-AAPK-pNA substrate rather than the anticipated increases (Wells et al.

00 WO 00/37658 PCT/US99/30362 S(1987) Proc. Nat. Acad. Sci. USA, 84: 1219-1223). Furthermore, while the G166D subtilisin BPN' mutant was reported to exhibit a 18-fold improvement in Arg/Phe P 1 selectivity this r was accompanied by a 2. 5-fold decrease in k.1/KM with the suc-AAPR-pNA substrate compared to WT (Ballinger et al. (196) Biochemistry 33: 13579-13585). Thus, compared to WT, both with respect to improved kat/KM with the suc-AAPR-pNA substrate and improved Arg/Phe Pi selectivity, the CMMs reported herein are more successful.

Furthermore, these improvements correlate directly with the number of introduced negative charges, such that each additional negative charge introduced at position 166 causes an approximate doubling in kt/KM with the complementary charged suc-AAPRpNA substrate (Figure 11). This arithmetic improvement is interesting since it contrasts the previous observation that while remote noninteracting charged mutations are additive, multiple interacting adjacent charged mutations often are not additive. Rather, in most cases the empirically observed effects are lower than predicted from their individual sums. so This phenomenon is particularly problematic for charged mutations due to the long range effects of electrostatic interactions. For example, the sum of the transition state stabilization energy, AAGIT, for the two single positively charged subtilisin mutations (D99K and E156K) over estimates the empirically observed effect of the double mutant, when assayed with an Arg PI containing substrate (Wells (1990) Biochemistry 29(37): 8509-8517). In contrast, in all cases the empirically determined AAGtT values for the aliphatic carbxylate series of mono- di- and tri- negatively charged side chains, of the S166C-S-h, and -j CMMs exhibit an additive effect of additional charge (AAGr RT ln[(kMt/KM)wT/(kat/KM)cMM] S166C-S-b -0.49; S166C-S-d -1.00; S166C-S-h -0.37; S166C-S-i -0.76; S166C-S-j -1.31 (kcal mol')).

Thus, the CMM approach offers a convenient method to circumvent the problem of the attenuation of the augmenting effect of the introduction of additional charges by SDM by permitting the introduction of a larger local charge density.

Conversely, the adopted design strategy of introducing a complementary positive charge in the Si binding cleft by the CMM approach to improve P 1 Glu selectivity was based on mimicking the specificity determinants of the serine proteases pronase (Nienaber et al. (1993) Biochemistry 32(43): 11469-11475; Svendsen et al. (1991) FEBS 292(1): 165-167) and granzyme B (Smyth et al. (1996) Leukoc. Biol. 60: 555-562; Murphy et al. (1998) Proteins: Structure: Function: and Genectics 4: 190-204; Caputo et al. (1994) Nature: Struct. Biol. 364-367), which exhibit a substrate preference for negatively WO 00/37658 PCT/IS99/30362 |charged Pi residues, and whose Si pockets are lined with positively charged residues. The success of the current approach is apparent from the remarkable 19-fold increase in kca/KM, c' with the suc-AAPE-pNA substrate displayed by S166C-S-CH 2

CH

2 NH3 This O enhancement is due to a combination of better binding, evident from the 2-fold lower KM, Z 5 and 8-fold higher kt (Table 3, entries 23, 24). The induction of electrostatic 00 complementarity was most unequivocally demonstrated by the 54-fold improvement in suc- AAPE-pNA to suc-AAPF-pNA substrate selectivity, with respect to kcaI/KM, for S 166C-S-

CH

2

CH

2 NH3 compared to WT. Previously, the individual G166R and G166K subtilisin BPN' mutations, elicited 23- and 340-fold improvements in kcJKM for the suc-AAPE-pNA ;10 substrate. 12'4 However, it must be noted that the El56Q-G166K double mutant was much more receptive to Glu P, and exhibited a 1900-fold improvement compared to WT (Wells et al. (1987) Proc. Nat. Acad. Sci. USA, 84: 1219-1223). Interestingly, both the G166R and SG166K mutants displayed even higher kc,/KMs with the hydrophobic PI residue containing substrates suc-AAPN-pNA and suc-AAPM-pNA and even with the positively charged Pi residue containing substrate suc-AAPK-pNA While S166C-S-CH 2

CH

2

NH

3 still exhibits a 10-fold preference for suc-AAPF-pNA compared to suc-AAPE-pNA (Table 3, entry 2) the substrate screen (Figure 10) shows that suc-AAPA-pNA and suc-AAPR-pNA are poorer substrates.

Since the S 166C-S-CH 2

CH

2 NH3 and suc-AAPE-pNA CMM-substrate pair exhibited the greatest knt/KM improvement relative to WT, at 19-fold (Table 3, entry 24), more detailed insights into the molecular basis of their interaction was sought using molecular modelling. Using the modelling approach reported previously (DeSantis et al.

(1998) Biochemistry 37: 5968-5973), the product inhibitor, AAPE bound to WT-SBL and to the S 166C-S-CH 2

CH

2 NH3 CMM was minimized. Molecular modeling revealed that J 25 the minimized binding conformations of AAPE to both the WT and S 166C-S-CH 2

CH

2

NH

3 enzymes are quite similar despite the 19-fold difference in kct/KM. However, the ammonium moiety of the S166C-S-CH 2

CH

2 NH3 CMM side chain is oriented toward the carboxylate of the glutamic acid PI residue, and although it is not quite within salt-bridge distance -to- -OOC, 4. 76 this additional favorable coulombic iriteraction between the ammonium side chain of S166C-S-CH 2

CH

2 NH3+ and the carboxylate of the glutamic acid P residue is deemed responsible for the observed 19-fold improvement in k.ct/KM for this CMM-substrate pair, compared to WT.

00 WO 00/37658 PCT/US99/30362 SConclusion: c The overall data clearly provides an encouraging validation of the practical effectiveness of the CMM strategy for generating complementary electrostatic and steric enzyme-substrate interactions. For each of the Ala, Arg, and Glu PI residues at least one, and up to five, of the designed CMMs exhibit improved kct/KMs compared to WT. The CMM approach is complementary to the SDM approach and also offers the additional opportunity for the introduction of multiply charged side-chains generating high charge densities at single active site locations. The beneficial effects of the introduction of a localized high charge density was demonstrated by the arithmetic increases in kt/KM, with 10 the suc-AAPR-pNA substrate, induced by incremental increases in the negative charge of the SI pocket.

Experimental: Sulfonatoethyl methanethiosulfonate (la) and ethylammonium methanethiosulfonate (Ib) were purchased from Toronto Research Chemicals (2 Brisbane Rd., Toronto, ON, Canada). Reagents Ic -f (Berglund etal. (1997) J. Am. Chem. Soc. 119: 5265-5266) and lh-lj were prepared as previously described. ES-MS data were acquired using a PE SCIEX API III Biomolecular mass spectrometer. The tetrapeptide substrates suc- AAPF/A/R/E-pNA were purchased from Bachem Bioscience Inc. (Torrance, CA). All buffer solutions were made up in deionized water.

Site-specific Chemical Modification To 25 mg of a S166C mutant, purified as previously described (DeSantis et al. (1998) Biochemistry 37: 5968-5973; Stabile et al. (1996) Bioorg. Med. Chem. Lett. 6(21): 2501-2506) and stored flash frozen in CHES buffer 5 mL; 70 mM CHES, 5 mM MES, 2 mM CaCl 2 pH 9. 5) at 20 OC was added one of the methanethiosulfonate reagents (la-g) (100 pL of a 0. 2 M solution), in a PEG (10,000) coated polypropylene test tube, and the mixture agitated in an end-over-end rotator. Blank reactions containing 100 pL of solvent instead of the reagent solution were run in parallel. Each of the modification reactions was monitored spectrophotometrically (E410 8800 cm (Bonneau et al. (1991) J. Am.

Chem. Soc. 113: 1026-1030) on a Perkin Elmer Lambda 2 spectrophotometer, by specific activity measurements. After the reaction was quenched by dilution in MES buffer (5 mM MES, 2 mM CaCI 2 pH 6. 5) at 0°C, the specific activity of the CMM (10 pL), was W 00/37658 PCT/US99/30362 Sdetermined in buffer containing: 0. 1 M TRIS pH 8. 6, 0. 005 Tween 80, and 1% DMSO, Swith the suc-AAPF-pNA substrate (1mg/mL) at 25 oC. The reaction was terminated when the addition of a further 100 pL of methanethiosulfonate solution effected no further change O in specific activity, generally in 30 min. to 3 h. The reaction solution was purified on a 0Z 5 disposable desalting column (Pharmacia Biotech PD-10, Sephadex G-25 M) pre-equilibrated 00 Swith MES buffer (5 mM MES, 2 mM CaCI 2 pH 6. 5) then dialyzed against 20 mM MES, 1 mM CaC 2 pH 5. 8 (3 x 1 L) at 4 *C and aliquoted into 0. 5-1. 5 mL volumes, flash frozen in liquid nitrogen and then stored at -20 C. Modified enzymes were analyzed by Snondenaturing gradient gels at pH 4. 2, run towards the cathode on the Pharmacia 10 Phast-System,TM and appeared as one single band.

Electrosprav Mass Spectrometry.

SPrior to ES-MS analysis, CMMs were purified by FPLC (BioRad, Biologic System) on a Source 15 RPC matrix (17-0727-20 from Pharmacia) with 5% acetonitrile, 0.

01% TFA as the running buffer and eluted with 80% acetonitrile, 0. 01% TFA in a one step gradient. Mass: WT: Calc. 26698, Found 26694. S166C-S-a: Calc. 26714, Found 26708 S166C-S-b: Calc. 26853, Found 26851. S166C-S-c: Calc. 26836, Found 26832. S166C- S-d; Calc. 26924, Found 26928. 70 S166C-S-e: Calc. 26886, Found 26890. S166C-S-f: Calc. 26842, Found 26844. S166C-S-g: Calc. 27128, Found 27123. S166C-S-h; Calc.

26846, Found 26846. 70 S 166C-S-i; Calc. 26890, Found 26894. 70 S 66C-S-i; Calc. 26934, Found 26939.

Regeneration of unmodified enzyme by treatment with B-mercaptoethanol.

To a solution of CMM 0 mg) in 250 pL of CHES-buffer (70 mM CHES, mM MES, 2 mM CaCI 2 pH 9. 5) was added 10 pL of a solution of P-mercaptoethanol (1 M in 95% EtOH). The reaction was monitored by specific activity measurements.

Free Thiol Titration The free thiol content of S166C CMMs, was determined spectrophotometrically by titration with Ellman's reagent (C412 13600 M cm') 8 7 in phosphate buffer 0. 25 M, pH 8. 0.

W000/37658 PCT/US99/30362 S iActive Site Titrations.

The active enzyme concentration was determined as previously described 7 by !monitoring fluoride release upon enzyme reaction with phenylmethanesulfonyl fluoride 0 (Aldrich Chemical Co. Inc. as measured by a fluoride ion sensitive electrode (Orion 0 5 Research 96-09). The active enzyme concentration determined in this way was used to calculate kinetic parameters for each CMM.

Kinetic Measurements Michaelis-Menten constants were measured at 25 °C by curve fitting (GraFit® 3. 03) of the initial rate data determined at eight concentrations 125 mM-8. 0 mM) of the suc-AAPX-pNA substrate in pH 8. 6 0. 1 M Tris-HCI buffer containing 0. 005% Tween 1% DMSO (E 41 0 8800 cm').

Molecular Modeling The X-ray structure ofsubtilisin Bacillus lentus 59 was used as the starting point for calculations on the wild type and chemically modified mutant enzymes. The enzyme setup was performed with Insight II. 88 To create initial coordinates for the minimization, hydrogens were added at the pH used for kinetic measurements. This protonated all Lys and Arg residues and the N-terminus and deprotonated all Glu and Asp residues and the C-terminus. In addition, the active site His64 was protonated. The model system with the Ala-Ala-Pro-Phe (from crystal structure) (Knapp et al. Brookhaven Database Entry 1JEA, [without Ala-Ala-Pro-Phe]) product inhibitor bound in the SI- S 4 pocket was solvated with a 5 A layer of water molecules giving a total number of water molecules of 1143 in this system. The overall charge of the enzyme-inhibitor complex resulting from this setup was +4 for the WT enzyme. Energy simulations were performed S with the Discover program (Discover [Biosym Technologies: Inc. San Diego: CA: USA]), on a Silicon Graphics Iris Indigo computer, using the consistent valence force field function (CVFF). A non-bonded cutoff distance of 18 A with a switching distance of 2 A was employed. The non-bonded pair list was updated every 20 cycles and a dielectric constant of I was used in all calculations. The WT enzyme was minimized in stages, with initially only the water molecules being allowed to move, then the water molecules and the amino acid side chains, and then the entire enzyme. The mutated and chemically modified enzymes were generated using the Builder module of Insight. Then the amino acid side chains within -71- -WO 00/37658 1 PCT/US99/30362 a 10 A radius of the a-carbon of the mutated residue were minimized while all other residues Swere constrained, then all of the atoms within a 10 A shell were minimized, followed by minimization of the whole system. To examine the effect of a different Pi residue (Glu), the O Phe to Glu mutation of the product inhibitor was constructed using insightII, and then this structure was minimized as above.

00 It is understood that the examples and embodiments described herein are for c illustrative purposes only and that various modifications or changes in light thereof will be suggested to persons skilled in the art and are to be included within the spirit and purview of S- 10 this application and scope of the appended claims. All publications, patents, and patent Sapplications cited herein are hereby incorporated by reference in their entirety for all 0 purposes.

-72-

Claims (16)

1. A modified enzyme wherein one or more amino acid residues in the enzyme are Sreplaced by cysteine residues, wherein the cysteine residues are modified by replacing the thiol hydrogen in said cysteine residues with a substituent group providing a thiol side chain comprising a multiply charged moiety. S2. A method of producing a modified enzyme, said method comprising: Sproviding an enzyme wherein one or more amino acids have been replaced with cysteine residues; and replacing the thiol hydrogen, in said one or more cysteine residues, with a C substituent group providing a thiol side chain comprising a multiply charged moiety.

3. An enzyme according to claim 1 or a method according to claim 2 wherein the enzyme is a serine hydrolase.

4. An enzyme or method according to claim 3 wherein the enzyme is a protease. An enzyme or method according to claim 4, wherein the protease is a subtilisin.

6. An enzyme or method according to claim 5, wherein said subtilisin is a Bacillus lentus subtilisin.

7. An enzyme according to claim 1 or a method according to claim 2, wherein the amino acid replaced with a cysteine is an amino acid selected from the group consisting of asparagine, leucine, methionine, and serine.

8. An enzyme according to claim 1 or a method according to claim 2, wherein the amino acid replaced with a cysteine is in a binding site of the enzyme.

9. An enzyme or method according to claim 8, wherein the amino acid is in a subsite selected from the group consisting of SI, Si', and S2. 00 74 O O An enzyme according to claim 1 or a method according to claim 2, wherein said enzyme is a subtilisin-type serine hydrolase and said cysteine is substituted for the Samino acid corresponding to a Bacillus lentus subtilisin residue selected from the group consisting of residue 156, reside 166, residue 217, residue 222, residue 62, residue 96, residue 104, residue 107, reside 189, and residue 209. _11. An enzyme according to claim 1 or a method according to claim 2, wherein said enzyme is a trypsinchymotrypsin-type serine protease and said cysteine is substituted N for the amino acid corresponding to a trypsin residue selected from the group consisting S 10 of Tyr94, Leu99, Glnl75, Aspl89, Serl90, and Glnl92.

12. An enzyme according to claim 1 or a method according to claim 2, wherein said enzyme is an alpha/beta serine hydrolase and said cysteine is substituted for the amino acid corresponding to a Candida antartica lipase (Protein Data Bank entry Itca) residue selected from the group consisting of TrplO4, Thrl38, Leul44, VallS4, Ile189, Ala 225, Leu278 and Ile185.

13. An enzyme according to claim 1 or a method according to claim 2, wherein the multiply charged moiety is negatively charged.

14. An enzyme or method according to claim 13, wherein the multiply charged moiety is selected from the group consisting of sulfonatoethyl thiol, 4-carboxybutyl thiol, 3,5dicarboxybenzyl thiol, 3,3-dicarboxybutyl thiol, and 3,3,4-tricarboxybutyl thiol. An enzyme or method according to claim 13, wherein the multiply charged moiety is a dendrimer or a polymer.

16. The enzyme of claim 1 or the method of claim 2, wherein the multiply charged moiety is positively charged.

17. An enzyme or method according to claim 16, wherein the multiply charged moiety is selected from the group consisting of aminoethyl thiol, 2- (trimethylammonium) ethyl thiol, 4,4-bis (aminomethyl)-3-oxo-hexyl thiol, and 2,2-bis (aminomethyl)-3-aminopropyl thiol. 00

18. An enzyme or method according to claim 16, wherein the multiply charged Smoiety is a dendrimer or a polymer.

19. A composition comprising an enzyme according to any one of claims 1 to 17 Cc and a detergent. A modified enzyme substantially as hereinbefore described with reference to any one of the figures, examples and/or preferred embodiments, excluding, if any, comparative examples.

21. A composition substantially as hereinbefore described with reference to any one of the figures, examples and/or preferred embodiments, excluding, if any, comparative examples.

22. A method of producing a modified enzyme substantially as hereinbefore described with reference to any one of the figures, examples and/or preferred embodiments, excluding, if any, comparative examples. Dated this thirty-first day of May 2007 Genencor International, Inc. Patent Attorneys for the Applicant: F B RICE CO