This invention relates to novel antibacterial agents and a method for predicting the activity thereof relative to penicillin. More particularly, this application describes a molecular modelling technique for determining the fit and reactivity of candidate compounds with bacterial cell wall receptors, and hence a method for predicting structural types 1hat exhibit activity.

It has been known since the 1940's that β-lactam antibiotics, such as the penicillins and cephalosporins, are effective by reason of their interference with the integrity of bacterial cell walls. It has also been discovered that the interference is effected by covalent bonding to the active site serine residue of one or more of a group of enzymes termed penicillin binding proteins (PBP's). These enzymes serve to complete bacterial cell wall synthesis by a cross linking of peptidoglycan chains, and are essential to the cells. All known PBP's include a sequence -Ser-X-X-Lys- and the simplest kinetic description of the reaction between a PBP and a β-lactam antibiotic is given in equation 1, below, where A is a generalized structure. Since the PBP is regenerated in the deacylation step, useful antibacterial activity is considered to require k ₃ / ≥ 1000 M ^"1 sec ^"1 and k ₄ ≤ 1 x 10 ^"4 sec ^"1 .

PBP-OH + Ji—M→ ?= FBF-OH-A ■► PBP-O-C H* — -*-> PBO-OH + ~ J i)

O £ e*_ _*-l*-t__±k_o ■cyiβUβn O dnβojr____U__π 0 OH

Step 1 Step 2

The question is, therefore, what is the correlation, if any, between antibacterial activity and the "lock-and-key" interactions which take place between the PBP and the antibiotic..

It is an object of the present invention to determine the correlation between antibacterial activity and the lock-and- key interactions between PBP's and selected antibiotics and thus provide a means by which the "fit" (Step 1) and "reactivity" (Step 2) of any selected candidate structure relative to the fit and reactivity of penicillin may be predicted with some degree of quantitative accuracy.

It is another object of this invention to design with this model novel non β-lactam compounds having antibacterial activity.

One aspect of the present invention provides a method of determining the molecular structure of large molecules wherein the strain energy of the molecule is minimized in terms of molecular parameters, characterized in that in order to identify starting parameters for the minimization procedure the one-point energies of a large number of most

probable random structures are first calculated, and a predetermined number of said random structures having the lowest energies are selected for said minimization procedure.

Thus by another aspect of this invention there is provided a method for determining fit and reactivity of any selected candidate antibacterial compound comprising (a) simulating the reaction of said compound with a model of a penicillin binding protein which includes a serine-lysine active site, by determining the relative ease of formation of a four- centred relationship between OH of said serine and a reactive site of said compound; and (b) determining the activation energy for the four-centred reaction of the chemically active functional group of said compound with methanol relative to the activation energy of the corresponding reaction of methanol with N-methylazetidinone.

Another aspect of this invention provides a non-β-lactam containing compound characterized in that said compound is capable of forming a four-centred transition structure which includes a serine OH group contained in a model of a penicillin binding protein, reacted therewith; said compound having an activation energy for reaction with methanol not greater than 3 kcal/mol higher than the activation energy exhibited by N-methyl- azetidinone.

Another aspect of this invention provides compounds of the formula:

where

X is selected from S, O, CH ₂ , H, R ₇ , and Se Y is selected from OH, NH ₂ , NHCOR ₉ , and SH R _1# R ₂ , R ₃ , R ₄ , R ₅ , R ₆ , R ₇ , are each hydrogen, alkyl, or aryl, and

R _g is a β-lactam active side chain, and pharmaceutically acceptable salts thereof, β-lactam active side chains are side chains known to be active in β-lactam antibiotics. As used herein, the substituents acceptable in beta-lactam antibiotics may be any of the wide range of permissible substituents disclosed in the literature pertaining to penicillin and cephalosporin compounds. Such substituents may, for example, comprise a group of the formula

-XQ

wherein X represents oxygen or sulfur and Q represents C^ alkyl (e.g., methyl or ethyl), C ₂ _ ₄ alkenyl (e.g. vinyl or propenyl) or aryl C^ alkyl (e.g., phenyl C^ alkyl such as benzyl) .

Such substituents also may be, for example, an unsaturated organic group, for example, a group of the formula

/

-CH=C \

R ²

wherein R^ and R ₂ which may be the same or different, and are each selected from hydrogen, carboxy, cyano, C ₂ . ₇ alkoxycarbonyl (e.g., methoxycarbonyl or ethoxycarbonyl) , and substituted or unsubstituted aliphatic (e.g., alkyl, preferably C^C _g alkyl such as methyl, ethyl, isopropyl or n-propyl) . Specific substituted vinyl groups of the above formula include 2-carboxyvinyl, 2-methoxycarbonylvinyl, 2- ethoxycarbonylvinyl and 2-cyanovinyl.

Alternatively, the β-lactam acceptable substituent may also be an unsubstituted or substituted methyl group depicted by the formula

-CH ₂ Y

wherein Y is a hydrogen atom or a nucleophilic atom or group, e.g., the residue of a nucleophile or a derivative of a residue of a nucleophile. Y may thus, for example, be derived from the wide range of nucleophilic substances characterized by possessing a nucleophilic nitrogen, carbon, sulfur or oxygen atom. Such nucleophiles have been widely

described in the patent and technical literature respecting β-lactam chemistry and are exemplified, for example, in Foxton et al U.S. Patent No. 4,385,177 granted May 24, 1983, at column 4, line 42 - column 8, line 24 and column 34, line 51 - column 36, line 17, the disclosure of which is incorporated by this reference herein.

Yet another aspect of this invention provides compounds of the formula:

where

X is selected from S, 0, CH ₂ , NH, NR ₈ , and Se

Y is selected from OH, NH ₂ , NHC0R _g , and SH

R.,, R ₂ , R ₃ , R ₄ , R ₅ , R _e , R ₇ , R ₈ are each hydrogen, alkyl, or aryl, and R _g is a a β-lactam active side chain, and pharmaceutically acceptable salts thereof

A further aspect of this invention provides compounds of the formula:

where

X-Y is selected from S-S, CH ₂ CH ₂ , S-CH ₂ , CH ₂ -S, S-NR ₈ , NR ₈ -S ,

CH ₂ H-0, 0-CH ₂ , 0-NR ₈ , NR _g -O, Se-Se, CH ₂ -CH ₂ , and Se-CH ₂

Z is selected from OH, NH ₂ , NHCOR ₉ , and SH

R R ₂ , R ₃ , R ₄ , R ₅ , R ₆ , R ₈ are each hydrogen, alkyl, aryl

R ₇ is alkyl, or aryl, and

R _g is a β-lactam active side chain, and pharmaceutically acceptable salts thereof

A still further aspect of the invention provides compounds of the formula:

where

X is selected from S, O, CH ₂ , NH, NR ₆ , and Se

Y is selected from N, CH, and CR ₇

Z is OH, NH ₂ , SH, or NHCOR ₉ (when Y=N)

Z is R ₁₀ (when Y=CH, or CR ₇ )

R ₁ =R ₂ =R ₃ =R ₄ =R ₅ =R ₆ =R ₇ = are each hydrogen, alkyl, or aryl, and

R _g is a β-lactam active side chain

^R ιo ^{= R _} CH -

where R ₁₁ is alkyl, or aryl, and

R ₁₂ =OH, NH ₂ , NHCOR _g , SH and pharmaceutically acceptable salts thereof.

Another aspect of the invention provides compounds of the formula:

where

X is selected from S, O, CH ₂ , NH, NR ₅ , and Se

Y is NR ₆ - Z, and

R ₁ R ₂ , R ₃ , R ₄ , R ₅ , and R ₆ are each H, alkyl, or aryl

Z is OH, SH, NH ₂ , or NHCOR ₇

R ₉ is a β-lactam active side chain, and pharmaceutically acceptable salts thereof. Preferably,

R ₆ is hydrogen and Z is NHCOR ₉ where R _g is lower alkyl and particularly benzyl.

As used herein, the term "alkyl" includes alkyl groups containing up to twenty carbon atoms, preferably C^ _g alkyl groups, which can optionally be monosubstituted, distributed or polysubstituted by functional groups, for example by free, etherified is esterified hydroxyl or mercapto groups, such as lower alkoxy or lower alkylthio; optionally substituted lower alkoxycarbonyloxy or lower alkanoyloxy; halogen; oxo; nitro; optionally substituted amino, for example lower alkylamino, di-lower alkylamino, lower alkyleneamino, oxo-lower alkyleneamino or aza-lower

alkylenearaino, as well as acylamino, such as lower alkanoylamino, lower alkoxycarbonylamino, halogeno-lower alkoxycarbonylamino, optionally substituted phenyl-lower alkoxycarbonylamino, optionally substitutedcarbamoylamino, ureidocarbonylaraino or guanidinocarbonylamino, and also sulfoamino which is optionally present in the form of a salt, such as in the form of an alkali metal salt, azido, or acyl, such as lower alkanoyl or benzoyl;

Optionally functionally modified carboxyl, such as carboxyl present in the form of a salt, esterified carboxyl, such as lower alkoxycarbonyl, optionally substituted carbamoyl, such as N-lower alkylcarbamoyl or N, N-di-lower alkylcarbamoyl and also optionally substituted ureidocarbonyl or guanidinocarbonyl; nitrile; optionally functionally modified sulfo, such as sulfamoyl or sulfo present in the form of a salt; or optionally O-monosubstituted or O, O-disubstituted phosphono, which may be substituted, for example, by optionally substituted lower alkyl, phenyl or phenyl-lower alkyl, it also being possible for 0-unsubstituted or 0- monosubstituted phosphono to be in the form of a salt, such as in the form of an alkali metal salt.

As used herein, the term "aryl" includes carbocyclic, hetrocyclic aryl. The carbocyclic aryl includes phenyl and naphthyl, optionally substituted with up to three halogen, C^ _g alkyl, C^ _g alkoxy, halo C^. _e ) alkyl, hydroxy, amino, carboxy, C^ _g alkoxycarbonyl, C _t _ ₆ alkoxycarbonyl-(C^ _g ) -alkyl, nitro, sulfonamido, C^g alkylcarbonyl, amido (-CONH ₂ ), or C ₆ alkylamino groups.

The term "heterocyclic" includes single or fused rings comprising up to four hetro atoms in the ring selected from

oxygen, nitrogen and sulphur and optionally substituted with up to three three halogen C ₃ . ₆ alkyl, C^ _g alkoxy, halo (C^. _g ) alkyl, hydroxy, amino, carboxy, C^ _g alkoxycarbonyl, C^.g alkoxycarbonyl (C^ _g ) alkyl, aryl, oxo, nitro, sulphonaraido, C^ _g alkyl-carbonyl, amido or C^ _g alkylamino groups.

Suitable C^ _g alkyl groups may be straight or branched chain and include methyl, ethyl n- or iso-propyl, n-, sec-, iso-, or tert-butyl. In those cases where the C^ _g alkyl group carries a substituent the preferred C^ _g alkyl groups include methyl, ethyl and n-propyl.

The invention will now be described in more detail, by way of example only, with reference to the accompanying drawings in which:

Figure 1 shows the structure of a model of a penicillin receptor whose docking to penicillins and cephalosporins leads uniformly to four-centered interactions between C-O-H of serine and (o) C-N of the penicillin or cephalosporin;

Figure 2 is a stereoscopic view of penicillin V docked to the peptide of Figure 1;

Figure 3 is a stereoscopic view of a A ³ -cephalosporin docked to the peptide of Figure 1;

Figure 4 is a stereoscopic view of a A ² -cephalosporin docked to the peptide of Figure 1;

Figure 5 is a stereoscopic view of a 4-epi-A ² - cephalosporin docked to the peptide of Figure 1;

Figure 6 is a close-up view of the four-centred interaction between C-O-H of serine and (O)C-N of the β-lactam ring which exists in Figure 2;

Figure 7 shows the N-protonated transition structure for the attack of methanol upon the exo face of N- methylazetidinone (ab initio calculation) ;

Figure 8 is the O-protonated transition structure for the attack of methanol upon the exo face of N- methylazetidinone (ab initio calculation) .

Figure ^' 9 is a stereoscopic view of the transition structure calculated using MINDO/3 for the reaction of methanol with a penicillin via an N-protonated pathway.

Figure 10 is a stereoscopic view of the transition structure calculated using MINDO/3 for the reaction of methanol with a penicillin via an O-protonated pathway;

Figure 11 is a stereoscopic view of the transition structure for the reaction of methanol with penam via endo-attack;

Figure 12 is a stereoscopic view of the complexation of 5 to the peptide of Figure 1; and

Figure 13 is a stereoscopic view showing the interaction of a cyclic structure with a model of a penicillin receptor.

Possible structures for peptides (e.g., enzymes),

penicillins and cephalosporins were examined using the computer program MMP2(85), which is available from the Quantum Chemistry Program Exchange (QCPE) at the University of Indiana, Blooraington, Indiana, U.S.A. This program calculates the strain energy of a molecule in terms of the contributions to this energy associated with stretching of bonds, bending of bond angles, torsion about bonds, and electrostatic and van der aals interactions of non-bonded atoms. To carry out the calculation, the Cartesian coordinates of all atoms must be entered, and lists of connected and attached atoms defined. If the types of atoms present in the molecule of interest are known, the strain energy can be minimized by the application of the Newton- Raphson procedure to an unconstrained multivariable non- linear function that includes all of the individual contributions noted above. This function is termed the force field. For the minimization to proceed in a reliable manner it is important that the geometry entered at the beginning of the calculation be reasonably accurate, and close to the bottom of an energy well.

For each different molecule to be examined with MMP2(85), it is first necessary to determine the parameters associated with the types of atoms present within this molecule. These parameters include, inter alia, standard bond lengths and bond angles, and stretching and bending force constants.

Bond lengths and angles are available from compilations of vibrational data, and others can be calculated by molecular orbital (MO) procedures. The general strategy for parameter development can be found in the monograph "Molecular Mechanics", by U. Burkert and N. L. Allinger, published by the American Chemical Society, Washington, 1982. Since the parameters for peptides (e.g., enzymes), penicillins and

cephalosporins to establish the force field required MMP2(85) were previously unknown, these were first determined and tested for their ability to reproduce known experimental crystal structures, and known effects of solvent upon the conformations (three-dimensional structures) of the different structural types. The parameters are termed PEPCON (Appendix 1) (for peptides), PENCON (Appendix 2) (for penicillins), and CEPARAM (Appendix 3) (for cephalosporins) .

A second necessary requirement for the use of MMP2(85) is the provision of the initial set of Cartesian coordinates. For small molecules, such as penicillins and cephalosporins, the coordinates of an experimental crystal structure can be used. Minimization with the appropriate parameters then leads to a calculated structure that reproduces the experimental structure. From this structure it is possible to proceed to other conformations and to the global minimum of the molecule by a series of dihedral drives around each of the dihedral angles of the molecule.

This is an option available in MMP2(85), and it works well. However, such a strategy is impractical for the analysis of a peptide because of the very large number of dihedral angles that would have to be examined for any such molecule which contains more than two or three amino acid residues.

Therefore, the computer programme ECEPP (Empirical

Conformational Energy Program for Peptides), which is available from QCPE, was modified to allow a random number generator to calculate the one-point energies of 200,000 initial structures containing permutations of the most probable backbone and dihedral angles. The fifty lowest

energy structures identified in this manner were read out, minimized using a quadratic minimization procedure, and then converted to MMP2(85) format for final minimization by the Newton-Raphson procedure. The objective of this initial search was to identify suitable starting parameters. This strategy has been tested extensively, works well, and has been applied to the treatment of a PBP, as described below.

More particulalry, in this procedure the strain energy of the molecule is minimized as a function of its dihedral angles with bond lengths and bond angles held constant. The minimization is preceded by a consideration of a subset of the parameters which form a basis for a specific subset of the complete parameter space, and the subset is comprised of the values 0, ±90, 180 degrees for the ø and >5 dihedral angles of the backbone, and the values -60 and 180 degrees for the first dihedral angle of the side chains. The w dihedral angle and all other side chain dihedral angles are maintained at 180 degrees. Each of the infinite number of points in this parametric subspace corresponds to an associated molecular strain energy. The subspace is then subjected to a sufficiently rich discrete randomly distributed uniform mapping so that there is an arbitrarily large probability that, some points (r) are found in a convex neighbourhood of local energy minima, and this set of points (r) is then used for the initialization of the minimization procedure. A reasonable number of points for the randomly chosen discrete subset described above is 200,000 in the case of a polypeptide containing up to 10 amino acid residues, and the set of points (r) preferably numbers 50.

With these procedures in place, an initial series of nine penicillins (la - li) was examined. Of these nine

compounds, la - Id are highly active antibiotics widely used in medicine (ampicillin, syncillin, penicillin G, penicillin V) , le - ^" If are significantly less active, and Ig - li are biologically inactive. The conformational analyses of these compounds revealed that antibacterial activity is associated specifically with a three dimensional structure in which the carboxyl group and side chain N-H project onto the convex face, and engage in hydrogen bonding lock-and-key interactions with the receptor, i.e., the PB .

CH- lb: R ~ PhO-C- I H

-17-

i£« R = PhCH ₂ C0-

1d ^; R _B Ph0CH ₂ C0-

H

I

1f R= P O— C- I CH-

Next a conformational analysis was performed on the cephalosporins 2a - 2c. Each of these has the phenoxyacetyl side chain, and can therefore be compared to penicillin V (Id) . The * ³ -isomer 2a is biologically active, but undergoes a facile equilibration with the - ^z - isomer 2b, which is biologically inactive. The reason for the lack of activity of 2b has not previously been established, but it has been suggested that the 4-epi-A ² - isomer 2c would exhibit a better fit to the PBP receptor, and possess antibacterial activity. However, such compounds are also inactive. The reason for this lack of activity is, therefore, also unknown.

Each of 2a.- 2c, like the penicillins la - Id, is found to prefer a conformation in which the side chain N-H occupies the convex face of the molecule. As with the penicillins, it can thus be postulated that lock-and-key interactions with the receptor involve primary binding by the carboxyl group and this side chain N-H.

CO*

H H

The active site serine D-alanyl carboxypeptidase- transpeptidase of Streptomyces R61 has been crystallized with incorporation of β-lactam compounds, and the crystal structure has been partially solved. The pH-dependence of the same enzyme has also been examined. Both kinds of studies suggest that the carboxyl group of a penicillin is closely associated with the protonated terminal amino group of the lysine residue of X-X-Lys. The crystal structure confirms that, in the complex, the β-lactam ring of penicillin is in close proximity to the active site serine. The pH-dependence study rules out involvement of a histidine residue in the chemical process, in contrast to the behaviour of chymotrypsin and related serine proteases. This result means that the serine O-H participates in the chemical reaction with the substrate.

These observations suggest that a valid model of the active site of a PBP can be obtained in terms of the amino acids that surround the unique serine residue, i.e., in this case, Val-Gly-Ser-Val-Thr-Lys.

Accordingly, the peptide Ac-Val-Gly-Ser-Val-Thr-Lys-NH- CH ₃ was subjected to an ECEPP search of 200,000 initial structures, followed by MMP2(85) refinement of 50 low energy structures identified in this search. One low energy structure having the lysine and serine side chains in proximity was found. This structure is characterized by the set of dihedral angles summarized in Table 1, and is shown as Figure 1.

Table 1

Dihedral angles of the model of the active site of the PBP of Streptomyces R61

The structure of Figure 1 has several features of interest. The convex face is mainly hydrophobic, and the concave face, which includes the serine and lysine side chains, is mainly hydrophilic. The concave face also contains the amide oxygen of the N-terminal acetyl group. These three sites are noted on Figure 1 as S (serine) , L (lysine) and A (acetyl) . The existence of a lock-and-key relationship between the concave face of Figure 1 and the previously determined convex face of penicillin and cephalosporin now seems clear. In terms of such a relationship, contact is required between the carboxyl group of the antibiotic and the terminal amino group of lysine, and also between the side chain N-H of the antibiotic and the acetyl oxygen.

The construction of a supermolecule in which the receptor is docked to a substrate through NH ₃ + ^" 0 ₂ C and N-H 0=C hydrogen bonds is, therefore, desirable. To obtain the structure and energy of such a supermolecule using the program MMP2(85), it is necessary to devise a procedure for the generation of a starting set of Cartesian coordinates.

A computer program has been written based on the following approach to the problem. Let A refer to a receptor molecule containing N, atoms, and B a substrate molecule containing N ₂ atoms, which is to be docked to A.

It is assumed that the geometries of A and B are known in Cartesian or internal coordinates, and that transformation between the two types of coordinate systems is possible. A start is thus made with (3^-6) and (3N ₂ -6) predetermined internal coordinates. To describe the geometry of the supermolecule containing (^ + N ₂ ) atoms requires 3(N, + N ₂ ) - 6 internal coordinates, i.e., six new internal coordinates must be determined and minimized. These comprise, typically, one bond length, two bond angles, and three dihedral angles, and they may be termed "intermolecular" internal coordinates.

To use the computer program, for which the source code listing is given in Appendix 4, one of the desired hydrogen bonding interactions is selected, and its distance set at 1.7-2.5 A, a typical intermolecular hydrogen bonding distance. Initial values are then given to the five remaining variables, and the energy is minimized, with the second hydrogen bond distance as a probe. The geometry of the resulting supermolecule, now expressed in Cartesian coordinates, is considered appropriate for MMP2(85) minimization when the second hydrogen bond distance is 1.7-2.5 A.

Figures 2-5 show stereoscopic views of the results of docking of the receptor model with, respectively, penicillin V, A ³ -cephalosporin V, ± ² -cephalosporin V and 4-epi-A ² - cephalosporin V. It can be seen that, in each case, the serine O-H sits on the convex face of the β- lactam compound, in such a manner as to create a four- centred interaction between 0-H and (O)C-N. This four- centred interaction is shown in closer detail for penicillin V in Figure 6.

From the Cartesian coordinates of C-O-H and (O)N-C of the optimized complexes, it is possible to compute the root

mean square deviations (rms) in A of the different four centred interactions, relative to a standard substrate, in this case penicillin V. When this is done for the series of penicillins la - li, it is found that all active penicillins have rms less than 0.2 A, and all inactive penicillins have rms greater than 0.4 A. For the series shown in Figures 2-5, the rms deviations are 0.000, 0.149, 0.338 and 0.148 A.

This implies that the "fits".of the biologically active A - ^3z -- cceepphhaalloossppoorriinn aanndd tthhee bbiioollooggiiccaallllyy iinnaaccttiivvee 44--eeppii-- - ² - cephalosporin to the penici Llllin receptor are identical. T Thhee biologically inactive A ² -cephalosporin has a poorer fit.

The biological activity of a drug depends not only on its ability to fit to a receptor, i.e., Step 1 of equation 1, but also on its ability to react chemically with the receptor, i.e., Step 2 of equation 1. The chemical reaction suggested by Figures 2-6 is a four centred process in which C7-0(Ser) (see A) and N-H(Ser) bond formation are concerted. This is an unprecedented chemical mechanism.

The hydrolysis and alcoholysis of β-lactam compounds has received much experimental and theoretical attention. In water above pH 8, the rate-determining step is addition to the carbonyl group to form a tetrahedral intermediate; below pH 6, there is rate-determining proton transfer to the β-lactam nitrogen, from the convex face of the molecule. Hydrolysis is extremely slow in the biologically relevant pH range 6-8, and the possible existence of a molecular (four-centred) mechanism in this region has not been established. Likewise, all previous theoretical studies of β-lactam hydrolysis have

emphasized anionic addition to the β-lactam carbonyl group.

Molecular orbital (MO) calculations of the ab initio type represent an accepted and well established procedure for the probing of the mechanisms of chemical reactions.

Such calculations can be performed using low level (STO- 3G) and high level (3-21G) basis sets using the computer programs GAUSSIAN 82 and GAUSSIAN 86, available from GAUSSIAN Inc., Pittsburgh, PA, U.S.A. Molecular orbital calculations of the semi-empirical type can be performed on relatively large molecular systems, and are valid once they have been calibrated with respect to an ab initio calculation' on the same system. The semi empirical procedures AMI, MNDO and MINDO/3 are available in the computer program AMPAC, available from QCPE.

Table 2 summarizes the ab initio data ( ±Ef, kcal/mol) for the reactions of N-methylazetidinone with water and with methanol via exo-oriented N- and 0- protonated structures. For the hydrolysis reactions, the 0- protonated structure is favoured by 1.75 kcal/mol at the lower ST0-3G level (STO- 3G//STO-3G) . One point calculations at the more appropriate 3-21G level (3- 21G//3-21G) increases the preference for the N-protonated transition structure to 5.66 kcal/mol. Analogous results are seen for methanolysis of N- methylazetidinone. These results prove that the four- centred interaction seen in Figures 2-6 reflects a genuine chemical process and, indeed, the energetically preferred chemical process. The N- and O-protonated methanolysis transition structures are shown in Figures 7 and 8, respectively.

Table 2 also summarizes the semi-empirical results for the hydrolysis and methanolysis of N-methylazetidinone, and it is evident that only MINDO/3 correctly reproduces

the preference for the N-protonated transition structure, Accordingly, MINDO/3 was used to examine the activation energies for the reactions of a large number of bicyclic azetidinones with methanol. These are summarized in Table 3.

Table 2

Relative *E for the Hydrolysis and Methanolysis of N- Methylazetidinone via N- and O-Protonated Transition

Relative energies are in kcal/mol.

Within each row of Table 3, the reactions of the different structural types are compared to that of the parent penam ring system of penicillin, and the data are discussed row-by-row:

Table 3

Calculated A±E (kcal/mol, MINDO/3) relative to N- Methylazetidinone for the Methanolysis of β-Lactam Compounds via Exo Formation of a Four-Centred N- Protonated Transition Structure

-2.80 -4..15 -4.11 -356

-2.80 -0.25 0. 0.46

-2 O -3.86 -2315

-2.80 -2.-V5 -2.93

-8 0 -2.73 -2.62

(1) the relative reactivities are carbapenam > penem > oxapenam > penam. Oxapenicillins and penems having the C3 and C6 substituents of penicillins are known to have antibacterial activity. Although the carbapenam ring system is known, carbapenicillins have not yet been prepared.

(2) in the comparison of the penam and cephera ring systems, the relative reactivities are penam > A ³ -cephem > * ² -cephem, acetoxymethyl- ± ³ -cephem. With a common acylamino side chain, penicillins are an order of magnitude more active than acetoxymethyl-* ³ -cephalosporins and the latter are, in general, an order of magnitude more active than 3-methyl-± ³ -cephems; ± ² -cephems are inactive.

(3) introduction of the C3 -carboxyl group enhances the reactivity. It is believed that the carboxyl group assists the methanolysis through hydrogen bonding, because epi ^' merization (C3β) decreases the reactivity significantly.

(4) introduction of C2-methyl substituents decreases the reactivity, unless a C3 -carboxyl group is present.

(5) the 6β-acylamino substituent has almost no effect on the reactivity. Consequently, the chemical reactivity of a penicillin differs only slightly from that of the parent penam.

Figures 9 to 11 show, respectively, stereoscopic views of the N- and O-protonated transition structures for exo- methanolysis of a penicillin- and O-protonated endo- methanolysis of penam. Such endo-oriented transition structures are ca 1 kcal/mol higher in energy than the O- protonated exo-structures and 5-6 kcal/mol higher in energy than the N-protonated exo-structures.

Table 4 summarizes the "fits" of penicillin V and 2a - 2c mentioned above, as well as the "reactivities" of the different ring systems, as given by ±±Ef for the reaction of methanol with the carboxylated substrates shown. The

product rms x ±±E represents a combination of fit and reactivity, and is seen to order correctly the different classes of antibiotics in the order of their biological activities. Based on this quantity, 2b is inactive because of its poorer fit to the receptor, and 2c is inactive because of its decreased reactivity.

The difference between 2b and 2c can be compared to the differences seen in Row 3 of Table 3. That difference is attributed to facilitation of the chemical process by hydrogen bonding of the attacking alcohol to the carboxyl

group when this group is on the convex face of the molecule. Thus 2c recovers the fit lost in 2b but concomitantly becomes less reactive. These considerations suggest that the attachment of a hydrogen bonding donor substituent on the convex face of 2c will restore the chemical reactivity while retaining the acceptable fit to the receptor. Possible sites for the attachment of the required substituent are sulfur, C4 and C7 (see Table 4 ^d for numbering) . Attachment of F, CH ₃ 0 and CH ₂ 0H to C4 and C7 in the required manner does not enhance the reactivity of 2c, but an alpha-oriented sulfoxide (3) exhibits reactivity superior to that of penicillin. Although a raalonic acid derivative which combines the favourable properties of 2b and 2c (4) exhibits somewhat reduced reactivity compared to penicillin ( ^A ±E = 3.51 kcal/mol), the product rms x A±Ef is intermediate between the active and inactive entries of Table 4. Accordingly, 3 and 4 are novel β-lactam containing structural types of potential biological interest.

Table 4

Root Mean Square (rms) Difference (A) , relative to Penicillin V, of the Cartesian Coordinates of the C-O-H

Atoms of Serine and the N-C=0 atoms of the Azetidinone Ring in the Complexes of β-Lactam Compounds with a Model of the Penicillin Receptor; Activation Energies (kcal/mol) for the Reaction of Azetidinones with Methanol, relative to the Penam Nucleus; and the Product

^a Refers to MINDO/3 calculations on

Refers to MINDO/3 calculations on

Refers to MINDO/3 calculations on

Refers to HINDO/3 calculations on

It is also possible to design entirely new structural types compatible with the combination of fit and reactivity developed here. Based on the dihedral angles of penicillin V, a carboxyl group oriented so that it makes a dihedral angle of 150-160° with a "reactive site", and a hydrogen bonding donor such as N-H or O-H oriented so that it makes a dihedral angle of -150 to -160° with the "reactive site" is required. The reactive site should be one that reacts with methanol via a four- centred transition structure, and with E no greater than 3-4 kcal/mol higher than that for the reaction with an azetidinone.

Systematic calculation of activation energies has identified the imino moiety (-C I=N-) as a functional group possessing the required reactivity, and incorporation of this moiety into a cyclic structure possessing dihedral angles of the required magnitude has identified structure 5 as a candidate structure having antibacterial activity by a penicillin-cephalosporin mechanism. The result is shown in Figure 12.

EXAMPLE 1

Application of PEPCON to the Calculation of the Polypeptide Crambin

This polypeptide contains 46 amino acid residues, 327 heavy atoms, and 636 atoms including hydrogens. The published crystal structure includes diffraction data refined to 1.5 A. The Cartesian coordinates of the heavy (non-hydrogen) atoms of this crystal structure were used as input to MMP2(85), hydrogens were added using an option available in MMP2(85), and Newton-Raphson minimization was performed using PEPCON. The calculated structure shows an rms deviation from the experimental structure of 0.291 A for the heavy atoms of the backbone, and 0.310 A for all heavy atoms.

EXAMPLE 2

Application of PENCON to the Calculation of Penicillin V

Repetition of the experiment of Example 1, with the Cartesian coordinates of the crystal structure of penicillin V and the PENCON parameters leads to an rms deviation of 0.1 A for all atoms.

EXAMPLE 3

Application of CEPARAM to the Calculation of Cephalosporin The Cartesian coordinates of the crystal structure of a - ^z -cephalosporin having the phenoxyacetyl side chain were entered, and the energy was minimized using MMP2(85) in conjunction with the CEPARAM parameters. The resulting rms deviation was 0.35 A.

^■ S I -

EX MPLE S

Application of the Random Number Strategy and ECEPP to the Conformational Analysis of a Peptide

The peptide Gly-Trp-Met-Asp-Phe-NH ₂ was entered into ECEPP, and an initial search was performed on 200,000 initial conformations of this molecule. The fifty lowest energy structures identified in this manner were minimized in ECEPP using a quadratic minimization procedure, and then refined using the PEPCON parameters of MMP2(85). One structure was strongly preferred, and the dihedral angles of this structure are identical to those of the gastrin tetrapeptide, which contains the Trp-Met-Asp-Phe-NH ₂ moiety of the above compound.

EXAMPLE 5

Calculation of the Structure of a Penicillin Receptor.

The peptide Ac-Val-Gly-Ser-Val-Thr-Lys-NHCH ₃ was treated as described in Example 4, and the fifty final structures were examined. Only one structure possessed lysine and serine side chains on the same side of the molecule. This structure is shown in Figure 1, and its dihedral angles are summarized in Table 1.

EXAMPLE 6

Dockinσ of Penicillin V to a Model of the Penicillin Receptor

The receptor model of Example 5 was docked to penicillin V using the computer program of Appendix 4. Several conformations of the penicillin were examined, and the final lowest energy complex is shown in Figure 2.

The compounds identified in this manner may thereafter be synthesized in accordance with standard chemical procedures known to persons skilled in the art.

The invention will be further illustrated by way of the following specific examples of compounds that have been prepared:

EXAMPLE 7: Synthesis of 3-Carboxy-5-Hydroxymethyl-6, 6- Dimethyl-± -l, 4-Thiazine

In formula I, X = S; Y = OH; R. = R ₂ = CH ₃ , R ₃ = R ₄ = R ₅ = R ₆ = H. Both D- and L- configurations at C ₃ are prepared.

STEP 1

Methyl isopropyl ketone (15 mL, 140 ramoles) was added to a solution of potassium chloride (1.1 g, 14.8 mmoles) in water (9.6 mL) . The mixture was stirred, warmed to 60 C, and illuminated with a 350 watt tungsten lamp mounted beside the flask. Bromine (11.9 g, 74.4 ramoles) was then added dropwise. When the colour of the first few drops had disappeared, the heating bath was replaced by a cold water bath, and the 350 watt bulb was replaced by a 60 watt bulb. Addition of bromine was continued at a rate sufficient to maintain the internal temperature at 40- 45 ^* C. When the addition was complete (25 min) the reaction mixture was allowed to stand for 2h and the organic phase was then separated, washed with water-

magnesium oxide and dried over anhydrous calcium chloride. Fractional distillation afforded 7 g of Al, b.p. 82-86°/145 torr. NMR (CDC1 ₃ ) 2.36 (3H, s), 1.77 (6H, s).

STEP 2

The bromeketone Al (4.65 g, 28 mmoles) was dissolved in glacial acetic acid (40ml), and freshly recrystallized lead tetraacetate (12.5 g, 28.2 mmoles) was added. The mixture was heated at 100°C, with stirring, for 2 h and cooled to room temperature. Ethylene glycol (2 mL) was then added to destroy unreacted lead tetraacetate. The reaction mixture was diluted with ether (100 mL) , washed successively with 10% sodium carbonate, water and saturated sodium chloride, dried and evaporated. The residue was distilled, and the fraction boiling at 57- 60°C/120 torr was further purified by chromatography (silica gel, 5% > 10% -> 15% ether-hexane) to give the bromoketoacetate Bl. NMR (CDC1 ₃ : 5.16 (2H, s), 2.13 (3H, s), 1.87 (6H, s).

STEP 3

Triethylamine (140 mL) was added to methylene chloride (3 mL) . The solution was cooled to -20°C, and gaseous hydrogen sulfide was introduced during 10 min. Then the bromoketoacetate Bl (200 mg) , in methylene chloride (1.0 mL) , was added dropwise with stirring during 10 min. The yellow solution was diluted with methylene chloride (30 mL) , washed successively with 2N hydrochloric acid, water and saturated sodium chloride, dried over anhydrous sodium sulfate and evaporated to yield the mercaptoketoacetate Cl. NMR (CDC1 ₃ ) 5.16 (2H, s) , 2.18 (3H, s), 1.57 (6H, s), 1.55(1H, s) .

STEP 4

To triphenylphosphine (258 mg, 0.98 ramole) in dry tetrahydrofuran ( 1 .0 mL) , at -78°C under a nitrogen

atmosphere, was added dropwise with stirring a solution of dimethylacetylenedicarboxylate (144 mg, 0.99 mmole) in tetrahydrofuran (1.0 mL) . The white slurry as maintained at -78°C for 10 min, and a solution of Boc-L (or D-)- serine (184 mg, 0.90 mole) in tetrahydrofuran (1.0 mL) was added dropwise. The temperature was maintained at - 78°C for 20 min and the reaction mixture was then allowed to warm to room temperature (2h) . The solvent was removed and the residue was chromatographed on silica gel. Elution with 15% -> 22% -> 30% -> 35% ethyl acetate-hexane afforded the beta-lactone DI. NMR (CDC1 ₃ ) 5.29 (IH, br) , 4.92 (IH, br) , 4.34 (2H, br) , 1.07 (9H, s).

STEP 5

To a solution of Cl (79.6 mg, 0.45 mmole) in dry degassed dimethylformamide (1.5 mL) was added dropwise a solution of lithium diisopropylamide (0.8 mmole) in tetrahydrofuran (1.5 mL) . The addition was carried out under nitrogen at -60°C. The reaction mixture was allowed to warm to -25°C during 50 min, cooled again to -55°C, and a solution of DI (56.4 mg, 0.30 mmole) in dry degassed dimethylformamide (0.5 mL) was added dropwise. When the addition was complete, the mixture was warmed to -20°C,

stirred for 25 rain and then diluted with ethyl acetate (30 mL) and washed with 0.5N hydrochloric acid (2 mL) . The aqueous layer was extracted with ethyl acetate (2 x 10 mL) and the combined organic extracts were washed with water (2 x 5 mL) and saturated sodium chloride (1 x 5 mL) , dried and evaporated. The oily residue was purified by preparative layer chromatography on a 10 x 20 cm plate coated with silica gel, using methylene chloride-ethyl acetate acetic acid (1.7:0.3:0.05) as eluant to give El (77 mg, 70.3%). NMR (CDC1 ₃ ) ^" 5.43 (IH, br) , 5.20 (IH, d, 18 Hz), 5.04 (IH, d, 18 Hz), 4.46 (IH, br) , 2.97 (IH, br), 2.78, 2.74 (IH, dd, 4.5, 9.0Hz), 2.17 (3H, s) , 1.48(3H, s), 1.47 (3H, S) , 1.44 (9H, s) .

STEP 6

The acid El (77mg) was dissolved in methylene chloride (10 mL) and treated at 0°C with an ethereal solution of diazomethane. The solvent was removed and the residue was purified on a 5 x 10 cm silica gel plate using hexane-ethyl acetate (1.4:0.6) as eluant to give the ester Fl (48. 2 mg) . NMR (CDC1 ₃ ) 5.32 (IH, br d) , 5.15 (IH, d, 11Hz), 5.07 (IH, d, 11Hz) , 4.48 (IH, br, q) , 3.76 (3H, s), 2.91 (IH, q, 4, 12Hz) , 2.74 (IH, q, 5.5, 12 Hz), 1.47 (6H, d), 1.44 (9H, S) .

Fi

STEP 7

The ester Fl (46 mg) , in tetrahydrofuran (1 mL) was treated at room temperature with 0.25 M lithium hydroxide (0.4 mL) . After 25 min an additional 0.4mL of lithium hydroxide was added. The mixture was stirred for 35 min and then diluted with ethyl acetate (10 mL) and washed with 0.5 N hydrochloric acid (2 x 5 mL) . The aqueous layer was extracted with ethyl acetate (2 x 5 mL) and the combined organic extracts were washed with water (1 x 5 mL) , followed by saturated sodium chloride (1 x 5 mL) , dried and evaporated. The residue was dissolved in the minimum of methylene chloride, treated with ethereal diazomethane, concentrated, and the residue was purified on a 10 x 20 cm silica gel plate. Elution with hexane- ethyl acetate (1.4 : 0.6) gave Gl (14.4 mg) . NMR (CDC1 ₃ ) 5.22 (IH, br), 4.58 (2H, d) , 4.48 (IH, br) , 3.75 (3H, s), 3.06 (IH, br) , 2.92 (IH, br) , 2.74 (IH, dd, 5, 11Hz), 1.46 (9H, s), 1.44 (6H, s) .

STEP 8

To a solution of Gl (5 mg, 0.015 mmole) in freshly dried pyridine (0.2 mL) were added successively silver nitrate (3.4 mg, 0.02 mmole) and t-butyldiphenylchlorosilane (6.3 mg, 0.023 mmole). The solution was stirred for 15 rain at room temperature under nitrogen. The solvent was then removed and the product was purified by preparative layer chromatography to give HI (5.5 mg) . NMR (CDC1 ₃ ) 7.69 (4H, m), 7.41 (6H, m), 5.07 (IH, br) , 4.70 (2H, s) , 4.41 (IH, br), 3.72 (3H, s), 2.70 (IH, dd) , 2.55(1H, dd) , 1.43(9H, s), 1.28(3H, S), 1.26(3H, s), 1.10 (9H, s) .

STEP 9

The silyated ester HI (5 mg) was treated at room temperature with formic acid (0.2 mL) . After 33 min the reaction mixture was frozen and the solvent was removed by lyophilization to yield the enamine II. NMR (CDC1 ₃ ):7.69 (4H, m) , 7.40 (6H, m) , 5.90 (IH, s) , 4.65 (IH, br) , 3.79 (3H, s), 3.76 (IH, br) , 3.17 (IH, dd, 10, 15Hz), 3.00(1H, dd,3,15 Hz), 1.49(3H, s), 1.31(3H, s), 1.08(9H, s).

STEP 10

The thiazine II was treated with lithium hydroxide, as described in Step 7, to remove the ester protecting group. The silylated protecting group was also removed in part to afford a reaction mixture which contained 3- carboxy-5-hydroxyraethyl-6,6-dimethyl ± -l,4-thiazine.

EXAMPLE 8: Synthesis of 3-Carboxy-5-(2-Hydroxypropyl) - 6,6-Dimethyl- ± ⁴ -l, -Thiazine

In formula II, X=S; Y=0H; R^R^CH _j ,- R ₃ =R=R ₅ =R ₆ =H; R ₇ =CH ₃ . Both D- and L- configuration at C3 are prepared, but the R- and S- epimers at C8 have not been separated; the D- isoraer is active.

^• 2-

STEP 1

A solution of ethyl 2-methylcyclopropanecarboxylate (5.0 g, 38.9 mmoles) in dry ether (5 mL) was added dropwise, with stirring under nitrogen, to the Grignard reagent prepared from magnesium turnings (1.935 g, 0.080 g-atom) and methyl iodide (12.43 g, 87.6 mmoles) in dry ether (42mL) . The addition required 30 rain; stirring was continued for 2.75 h at room temperature and then for 2 h under reflux. The reaction mixture was cooled in an ice- bath and saturated ammonium chloride (lOmL) was added, with stirring. The layers were separated and the aqueous layer was extracted with ether (2 x 20raL) . The combined organic phase was dried, evaporated and the residue distilled at 132-136°C to give the tertiary alcohol A2 (4.24 g, 95%)

STEP 2

To the alcohol A2 (4.24 g, 37 mmoles) cooled in an ice- bath, was added ice-cold 48% hydrobromic acid (15 mL) . The mixture was shaken vigorously in the ice-bath for 30 min. The two layers were then separated, the aqueous layer extracted with hexane (2 x 20 mL) , and the combined organic phase was washed successively with saturated bicarbonate (2 x 10 mL) , water (2 x 10 mL) and saturated sodium chloride (2 x 10 mL) , dried over anhydrous sodium sulfate, and evaporated. Distillation afforded 3.72 g (60%) of the bromide B2, b.p. 46-54°C/10 torr.

STEP 3 To a solution of the bromide B2 (3.72 g, 21 mmoles) in glacial acetic acid (20mL) was added potassium acetate (3.1g, 31.6 mmoles). The mixture was heated under reflux for 12 h, cooled, and poured into water (30mL) . Extraction with ether (3 x 30 mL) , followed by successive

washing of the organic phase with saturated sodium carbonate, water and saturated sodium chloride, drying, and evaporation at room temperature yielded the acetate C2, 2.82 g (85%). NMR CDC1 ₃ ) 5.10 (IH, brt) , 4.88 (IH, q, 6Hz), 2.30 (IH, m), 2.19 (IH, m) , 2.02 (3H, s) , 1.71 (3H, br s), 1.62 (3H, br s), 8.00 (3H, d, 6Hz) .

Step 4

The acetate C2 (320 mg, 2.05 mmoles) was dissolved in methanol (2raL) and treated dropwise with a 1.5 M solution of potassium hydroxide in methanol (1.38 mL) . The reaction mixture was allowed to stand for 6h and was then neutralized with 1.5 M methanolic hydrogen chloride, and the solvent was removed. The residue was dissolved in methylene chloride, and this solution was washed successively with water and saturated sodium chloride, dried and evaporated to give the alcohol D2 (208 mg, 99%) .

STEP 5A

The alcohol D2 (312 mg, 2.73 mmoles) was dissolved in dimethylforraaraide (2mL) and to this solution were added successively t-butyl dimethylchlorosilane (535mg, 3.55 mmoles) . The mixture was stirred for 2h and then filtered. The insoluble material was triturated with ether (20mL) and the combined organic material was washed successively with saturated sodium bicarbonate, water and saturated sodium chloride, dried and evaporated to give the silyated compound E2A (620 mg, 100%) .

STEP 5B

The alcohol D2 (25mg, 0.22 mmole) was dissolved in dimethylformaraide (0.2mL), and the solution was treated successively with pyridine (27 μl, 0.33 mole), t- butyldiphenylchlorosilane (90 μL, 0.35 mmole) and silver nitrate (56mg, 0.33 mmole) . The mixture was stirred at room temperature for 4 h, and the product was then isolated, as described in Step 5A, to yield E2B.

44

STEP 6A

The olefin E2A (624mg, 2.73 ramoles) was dissolved in acetone (3mL) and 18-crown-6 (lOOmg, 0.27 mmole) and acetic acid (0.16mL) were added successively followed, dropwise, by a solution of potassium permanganate (603mg, 3.82 mmoles) in water (7.5mL). The mixture was stirred for 1 hr and then diluted with methylene chloride (50mL) . The organic phase was washed successively with 20% sodium bisulfite, 0.5 N hydrochloric acid, saturated sodium bicarbonate, water and saturated sodium chloride, dried and evaporated. The residue was subjected to flash chromatography on silica gel (7g) . Elution with 4 -> 15% ethyl acetate-hexane gave 479 mg (70%) of the ketol F2A.

STEP 6B

The olefin E2B (77.5mg, 0.22 mole) was oxidized with potassium permanganate, as described in Step 6A, to yield the ketol F2B. NMR (CDC1 ₃ ) : 7.72 (4H, m) , 7.43 (6H, m) ,

4.43 (IH, q, 6Hz) , 3.81 (1H,S), 2.81 (IH, dd, 5, 16Hz),

2.58 (IH, dd, 7, 16Hz), 1.31 (3H, s) , 1.29 (3H, s), 1.10

(3H, d, 5Hz), 1.04 (9H, s)

STEP 7A To a solution of the ketol F2A (478 mg, 1.83 mmoles) in methylene chloride (6mL) were added successively triethylamine (0.76 mL, 4.0 mmoles) and methanesulfonyl chloride (0.24mL, 3.1 mmoles) . The reaction mixture was stirred for 5h at room temperature and then diluted with methylene chloride (80mL) . The solution was washed successively with water, 0.5 N hydrochloric acid, saturated sodium bicarbonate, water and saturated sodium chloride, dried and evaporated. Flash chromatography on silica gel (3g) and elution with 7% -> 8% -> 9% -> 10% ethyl acetate-hexane gave G2A (432 mg, 70%) .

-46 -

STEP 7B

The ketol F2B (277mg, 0.72 mmole) was converted into the mesylate G2B (233mg) , as described in Step 7A. NMR (CDC1 ₃ ) 7.71 (4H, m) , 7.41 (6H, m) , 4.44 (IH, dd) , 3.08 (3H, s), 2.95 (IH, dd, 6, 18Hz) , 2.27 (IH, dd, 7, 18Hz), 1.63 (3H, s), 1.61 (3H, S) , 1.15 (3H, d, 6Hz) , 1.06 (9H, s )

STEP 8A

Methylene chloride (5mL) was saturated with hydrogen sulfide at -20°C, and triethylamine (0.14 mL, 1 mmole) and a solution of the mesylate G2A (233mg, 0.5 mmole) were added successively. The solution was stirred for 10 min at -20°C and for 45 rain at -20°C -> 0°C, and was then diluted with methylene chloride (30mL) , washed successively with 0.5 N hydrochloric acid, water and

saturated sodium chloride, dried and evaporated to give, after drying at 0.1 torr, the mercaptan H2A (170mg, 85%). NMR(CDC1 ₃ ) 4.37 (IH, m) , 2.98 (IH, dd, 5, 11Hz), 2.63 (IH, dd, 4, 11Hz), 1.98 (IH, s), 1.49 (3H, s) , 1.48 (3H, s), 1.17 (3H, d, 5Hz), 0.84 (9H, s), 0.05 (3H, s), 0.01 (3H, s)

^■ b

STEP 8B

The mesylate G2B was converted into the mercaptan H2B as described in Step 8A. NMR (CDC1 ₃ ) 7.71 (4H, m) , 7.40 (6H, m), 3.00 (IH, dd, 6, 16Hz), 2.75 (IH, dd, 7, 16Hz), 1.93 (IH, s), 1.46 (3H, s), 1.45 (3H, s) , 1.13 (3H, d, 6Hz) , 1.05 (9H, s)

STEP 9

Under nitrogen, the raercaptan H2A ( lOOmg, 0.36 mole ) was dissolved in degassed dimethylorraamide ( 1 .0 mL) . The

solution was cooled to -55°C and treated with 0.45raL of a solution of lithium diisopropylamide prepared from n- butyllithiura (0.8mL of a 1.6M hexane solution) and diisopropylamine (0.36mL, 0.259g, 2.56 ramoles) in degassed tetrahydrofuran (0.8mL). The reaction mixture was stirred at -45°C for 30 min, and a solution of the beta-lactone DI (D- or L) (56.8mg, 0.30 mole) in degassed dimethylformaraide (O.δraL) was added. The mixture was stirred at -30°C for 20 min and then diluted with methylene chloride (lOmL) and washed with 0.5N hydrochloric acid. The aqueous layer was extracted with methylene chloride (2 x 5mL) and the combined organic extracts were washed with water, then saturated sodium chloride, dried and evaporated. The residue was dried under high vacuum and purified by flash chromatography (silica gel, 4g; 0% -> 8% ethyl acetate-methylene chloride (1% acetic acid)) to give the coupled product I2D or I2L I2D (88.6%, [ ] _D -2.27 (c 0.1, chloroform)). NMR (CDC1 ₃ ) (one isomer) 5.28 (IH, br t) , 4.48 (IH, br) 4.32 (IH, ra), 2.83, 2.71 (2H, m) , 2.71, 2.62 (2H, ra) ,

1.44 (9H, s), 1.43 (6H, s), 1.16 (3H, d, 6Hz) , 0.85 (9H, ), 0.05 (3H, s) , 0.00 (3H, s) . The nmr spectrum shows a 1:1 mixture of epimers in the 2-hydrox propyl side _. chain.

I2L ( 83% , [ ] _D +2 .33 (c 0. 1 , chloroform) ) . NMR (CDC1 ₃ ) (one isomer) 5 .28 ( IH, br t) , 4.48 ( IH, br) , 4.32 ( IH,

m) , 2.86, 2.79 (2H, ra) , 2.70, 2.61 (2H, m) , 1.43 (9H, s), 1.42 (6H, s), 1.16 (3H, d, 6Hz) , 0.83 (9H,s), 0.04 (3H, s), 0.00 (3H, s) . The nmr spectrum shows a 1:1 mixture of epimers in the 2-hydroxypropyl side chain.

STEP 10A

To I2D (22.7 mg, 0.049 mmole) was added formic acid (0.3raL) . The solution was shaken for 20min at room temperature and the solvent was then removed by lyophilization. The residue was dissolved in a mixture of ether (3 mL) and water (lmL). The ether phase was extracted with water (lmL), and the combined aqueous phase was neutralized with 5% sodium bicarbonate and lyophilized to give 2 (5mg, 40%) having the D- configuration at C3, as a mixture of epimers in the 2- hydroxypropyl side chain. NMR (D20) 4.23 (IH, m) , 3.80 (IH, m), 3.30 (IH, q) , 2.70-2.85 (3H, ra) , 1.40 (6H, s) , 1.15(3H, d) .

^• 50-

STEP 10B

The procedure of Step 10A was repeated on 12L to give 2 having the L-configuration at C3

_-____ ^~ __ ^e e ^a c* ^M 2. —

EXAMPLE 9 Bioassy of 2.-D

The compound was assayed for antibacterial activity on plates inoculated either with Sarcina lutea or Escherichia coli. In the former case, penicillin G was employed as a standard. In the latter case, Cephalexin was employed as the standard. The compound was found to be 800 times less active than penicillin G, and 10 times less active than Cephalexin. The L-isomer of 2. was found to be inactive in both assays.

EXAMPLE 10 Synthesis of 2-Thia-4-Carboxy-6- (2- Hydroxypropyl) -7,7-Dimethyl- 5-1,5-Thiazepine.

^■ 51 -

In formula III, X-Y » S-S; Z ■ OH; R ₁ =R ₂ =R ₇ =CH ₃ ; R ₃ =R ₄ =R ₅ =R ₆ . Both D- and L- configurations of C4 are prepared, but the R- and S - isomers at C9 have not been separated. The L-isoraer is active (figure 13)

L-Cysteine hydrochloride (4.1mg, 0.026 mmole) was dissolved in 90% methanol-water (0.35mL), and a solution of the mercaptan H2A (Example 2, Step 8A) (7.1 mg, 0.026 mmole) in methanol (0.35mL) was added, followed by iodine (6.5mg, 0.026 mmole) and triethylamine (7 μL, 0.050 mmole) . The reaction mixture was left for 30 rain at room temperature and the solvent was then removed under reduced pressure. The residue was partitioned between pH 7 phosphate buffer (containing one drop of 10% sodium thiosulfate) and methylene chloride. The aqueous layer was extracted with ethyl acetate (1 x 5 mL) and lyophilized. The residue was triturated with methanol, and the methanol extract was combined with the methylene chloride and ethyl acetate extracts and evaporated. The product was purified on a 10 x 15 cm alumina plate using methylene chloride-methanol-water (1.8 : 0.2 : 0.15) as eluant to give the disulfide A4-L (8.9mg, 80%). NMR (D ₂ 0) : 4.19 (IH, m) , 3.92 (IH, dd, 3.7 Hz), 3.18 (IH, ra) , 3.04 (IH, ra), 2.92 (IH, m) , 2.76 (IH, m) , 1.43 (6H, s), 1.12 (3H, d, 7Hz) . The compound is a mixture of epimers in the 2-hydroxypropyl chain.

Repetition of this experiment using D-cysteine in place of L-cysteine gave A4-D.

The acids A4-D and A4-L were dissolved in water containing sodium bicarbonate and assayed for antibacterial activity by plate assay using S. lutea. A zone of inhibition was observed with the L-isomer, but not with the D-isomer. The inhibition is ascribed to the formation of the cyclic structure 3L, whose interaction with the model of the penicillin receptor is shown in Figure 13.

EXAMPLE 11: Synthesis of 3-Carboxy-5-Oximino-1,4-Thiazine In formula IV, X=S; R^R^H; R ₃ =R ₄ =R _g =H; X=N; Z =OH. Both D- and L- isoraers are described.

STEP 1

To a solution of D-cysteine (605.8mg, 5 ramoles) in methanol (lOmL) were added successively ethyl broraoacetate (0.99g, 5.95 ramoles) and triethylamine

(1.4mL, 1.02g, 10 ramoles). The solution was stirred for 20 min at room temperature and ether (20mL) was then added. The product was collected by filtration, washed with ether and dried. Five hundred mg of this material were suspended in dimethylforraamide (5mL) , and p- toluenesulfonic acid (458mg, 2.41 ramoles) was added. The resulting solution was treated portionwise with diphenyldiazomethane until the color of the diazo compound persisted, and the reaction mixture was stirred overnight. It was then diluted with ether (20mL) and extracted with water (2 x lOmL) . The aqueous extract was made alkaline by addition of saturated sodium carbonate, and was then extracted with ethyl acetate (3 x lOmL) . The combined organic extracts were washed with water and saturated sodium chloride, dried and evaporated. A 370- mg portion of the residue (0.99 mmole) was dissolved in 1,4-dioxane (8mL), 2-pyridone (47 mg, 0.49 mole) was added, and the solution was heated under nitrogen at 102°C for 7 h. Additional 2-pyridone (23.5 mg, 0.25 mmole) was then added and heating was continued for 4 h. At this time the solvent was removed under reduced

pressure and the residue was purified on 15 g of silica gel. Elution with 8% ethyl acetate-hexane afforded 252 mg (78%) of the thiazinone benzhydryl ester A5-D. NMR (CDC1 ₃ ) 7.34 (10H, m) , 6.96 (IH, s), 6.48 (IH, s) , 4.46 (IH, ra) , 3.33 (2H, s) , 3.21 (IH, dd, 4, 15Hz) , 2.98 (IH, dd, 9, 15Hz) .

f£-2

STEP 2

The thiazinone ester A5-D (252 mg, 0.77 mmole) was dissolved in dry tetrahydrofuran (5mL) under nitrogen, and the reagent prepared from phosphorous pentasulfide and diphenyl ether according to Tetrahedron Letters 3815 (1983) (244rag, 0.46 mole) was added. The solution was stirred for 35 min, concentrated, and the residue was purified on silica gel (8g) . Elution with 15% ethyl acetate-hexane afforded 214 mg (81%) of the thioamide B5- D. NMR (CDCl _j ) 8.59 (IH, S), 7.35 (10H, m) , 6.98 (IH, s), 4.39 (IH, ra), 3.79 (2H, s) , 3.32 (IH, dd, 4, 15Hz), 3.02 (IH, dd, 8, 15Hz) .

STEP 3

The thioamide B5-D (80mg, 0.23 mmole) was dissolved with stirring in ice-cold dry tetrahydrofuran (92mL) under nitrogen and sodium hydride (80%, 8.4mg, 0.28 mmole) was added. After 5 min stirring in an ice-bath, the reaction mixture was treated with 30 μL (0.48 mmole) of methyl iodide. Reaction was complete after 25 min. Dilution with ether, followed by successive extraction with water, saturated sodium bicarbonate and saturated sodium chloride, drying and evaporation gave a product which was purified on silica gel (3g) . Elution with 10% ethyl acetate-hexane afforded 59.1mg (75%) of the thiomethylimine C5-D. NMR (CDC1 ₃ ) 7.35 (10H, m), 6.96 (IH, s), 4.53 (IH, ra) , 3.27 (IH, dd, 5, 18Hz) , 3.15 (IH, dd, 5, 18Hz), 2.99 (IH, dd, 3, 13Hz), 2.81 (IH, dd, 4, 13Hz), 2.37 (3H, S) .

STEP 4

The thiomethylimine C5-D (59 mg, 0.165 mole) was dissolved in tetrahydrofuran (0.5mL) and added to a solution prepared under nitrogen from hydroxylamine hydrochloride (68.8mg, 0.99 mmole) and 1.65 M methanolic sodium methylate (0.3 mL, 0.5 mmole) in methanol (0.7mL). The reaction was complete in 10 rain. The mixture was diluted with methylene chloride (lOmL), washed successively with saturated sodium bicarbonate, water and saturated sodium chloride, dried and evaporated. Chromatography on silica gel (1.5g) and elution with 12% ethyl acetate-methylene chloride gave 52.7 mg (94%) of the oximino ester D5-D. NMR (CDC1 ₃ ) 7.34 (11H, ra) , 6.93. (IH, s), 5.97 (IH, s) , 4.28 (IH, ra) , 3.30 (IH, d, 13Hz) , 3.21 (IH, dd, 3, 13Hz), 3.16 (IH, d, 13Hz) , 3.08 (IH, dd, 7, 13Hz) .

STEP 5

The ester D5-D (47 mg) was dissolved in formic acid (lmL). After 5h at room temperature the reaction mixture was frozen and the solvent removed by lyophilization.

The residue was partitioned between ether and water, the ether layer was extracted once with water, and the combined aqueous extracts were lyophilized again to yield 4-D. NMR (D ₂ 0) : 4.15 (IH, m) , 3.50 (IH, d, 14Hz) , 3.34 (IH, d, 14Hz), 3.15 (IH, dd, 6, 15Hz) , 3.02 (IH, dd, 6,

15 Hz)

The L-enantioraer of 4 was prepared as described above, but starting with L-cysteine in place of D-cysteine .

-58 -

Antibacterial activity was observed on the D-isomer.

EXAMPLE 12: Synthesis of 3D-Carboxy-5- Phenylacetylhydrazil-A -τhiazine.

A solution of thioraethylimine C5-D (11 mg, 0.031 mmole) and phenylacetic hydrazide (9.2 mg, 0.062 mmole) was stirred overnight under nitrogen in methylene chloride (0.6 mL) . The reaction mixture was purified by preparative layer chromatography on silica gel to give the adduct P (14mg, 98%). NMR (CDCL ₃ ) : 7.34 (IH, m), 6.89 (IH, s), 6.55 (IH, br S) , 4.24 (IH, br s), 3.78 (2H, s) 3.55 (IH, br, s) , 3.33 (IH, d, 15 Hz), 3.17 (IH, d, 15 Hz), 3.05 (2H, br) .

Ϋ

-59 -

The adduct P (10 mg, 0.022 mmole) was treated with formic acid (0.4 raL). The solution was allowed to stand at room temperature for 5 h and the solvent was then removed by lyophilization. The residue was partitioned between ether (0.2 mL) and water (0.2 mL) . The ether layer was extracted once with water (0.2 mL) , and then the combined aqueous phase was freeze dried to give the product Q. (3 mg, 47%). NMR (D ₂ 0, NaHC0 ₃ ) : 8.33 (IH, s) , 7.28 (5H, m), 3.98 (IH, ra) , 3.55 (2H, s) 3.35 (IH, d, 17.5 Hz), 3.12 (IH, d, 17.5 Hz), 3.11 (IH, br d, 15 Hz), 2.83 (IH, br d, 15 Hz) , 2.83 (IH, s) .

The L-isomer QL was prepared in the same way, starting with C5-L.

.

5 -0.058 6 0.810 8 4.524 19 3.260 27 -0.700 5 0.0 6 0.000 7 3.010 9 3.320 11 3.950 12 1.850 21 -1.960 4 0.000 5 0.000 26 2.120 27 1.120 23 -1.410 14 -1.810 24 0.000 23. -1.350 15 0.000 25 0.000 23 -2.270 26 1.430 27 -1.030 27 0.000

4 0.044 1.940

11 0.066 1.780

12 0.050 1.740

13 0.030 1.900

14 0.017 0.930 19 0.055 1.820 5 0.036 1.250 6 0.055 1.820 7 0.044 1.940 - 1 0.45 110.30 1 1 1 0.45 111.20 2 1 1 0.45 112.40 3 1 2 0.58 114.00 1 3 0.67 107.80 1 - 3 0.67 110.80 2 1 3 0.67 112.20 3 1 4 0.71 113.10 1 5 0.36 109.39 1 6 0.56 109.10 1 1 6 0.56 104.10 2 1 6 0.56 109.40 3 1 8 0.57 109.47 1 9 0.56 109.40 1 - 9 0.56 109.60 2 1 9 0.85 111.10 3 1 13 0.90 111.20 1 15 0.63 108.80 19 0.75 111.20

SUBSTITUTE SHEET

^• 69

APPENDIX 4 PROGRAMME FOR DOCKING OF TWO MOLECULES

PARAMETER (NT=150)

PARAMETER (NG=300)

CHARACTER*2 ASYM , ITLE

INTEGER TYPEA,TYPEB ,TYP

COMMON/COOD / COORD(3,NG) ,CHARGE(NG)

COMMON/TYPE / TYPEA(NT) ,TYPEB(NT) ,TYP(NG)

COMMON/SYMM / ASYM(NG) ,TITLE(40)

COMMON/PARAM/ VEP(NG) ,VRA(NG)

COMMON/CORD / XA(NT) ,YA(NT) ,ZA(NT) ,XB(NT) ,YB

COMMON/FINAL/ TXB(NT) ,TYB(NT) ,TZB(NT) ,CA(NT)

COMMON/INFO / NA,NB,IP1,IP2,IP3,IP4,R1,R2,Cl

COMMON/PATH/MYWAY

DIMENSION X(5) ,E(5)

DATA E /5*0.01/

READ(8, 10 ) TITLE 10 FORMAT(40A2)

READ(8,*) IP1,IP2,IP3,IP4 ,MYWAY,Rl,R2,SC C MYWAY=1 : DISTANCE C MYWAY=2 : ENERGY C SC = SPECIAL CHARGE

CALL COMBIN(NA,NB)

READ(8,*) THETA,PI,XROT,YROT,ZROT ,SCALE,JCO

E(l) = EE1

E(2) = EE1

E(3) = EE1

E(4) = EE1

E(5) = EE1 C READ(5,*) (E(I),I=1,5) C SPECIAL CHARGE FOR IP2 AND IP4

CIP2= SC

CIP4= -SC

IPRINT=1

ICON = 1 C CONVERT DEGREE TO RADIAN

DEGREE=57.29577951D0

X(l)= THETA/DEGREE

X(2)= PI/DEGREE

X(3)= XROT/DEGREE

X(4)= YROT/DEGREE

X(5)= ZROT/DEGREE

NVAR=5 C C READ COORDINATE WITH MM FORMAT

READ(4,20) (XA(I),YA(I),ZA(I),TYPEA(I),I=1,N 20 FORMAT(2(3F10.5,I5,5X))

READ(5,20) (XB(I) ,YB(I) ,ZB(I) ,TYPEB(I) ,1=1,N ^' CALL SYMBOL

CALL IWRITE

CALL PARM

CALL CHARG

SUBSTITUTE SHEET

WRITE(6,22) (CHARGE(I),I=1,12) 22 FORMAT(3X,6F10.4) C

CALL OPTIM(X, VAR,SCALE,IPRINT,ICON,E) C C CONVERT RADIAN TO DEGREE

DO 30 J=l,5 30 X(J)= X(J)*DEGREE

WRITE(6,35) X 35 FORMAT(4X, ' OPTIMIZED THETA-PI-X-Y-Z ANGLES C WRITE BOND DISTANCE BETWEEN IP2 AND IP4

X1=XA(IP2)

Y1=YA(IP2)

Z1=ZA(IP2)

X2=TXB(IP4)

Y2=TYB(IP4)

Z2=TZB(IP4)

R12=DIST(X1,Y1,Z1,X2,Y2,Z2)

WRITE(6,40) IP2 ,IP4 ,R12

DO 50 K=1,NA

COORD(1,K)=XA(K)

COORD(2,K)=YA(K)

COORD(3,K)=ZA(K) 50 CONTINUE

NTOT=NA+NB

DO 60 KK=1,NB

COORD(1,NA+KK)=TXB(KK)

COORD(2,NA+KK)=TYB(KK)

COORD(3,NA+KK)=TZB(KK) 60 CONTINUE

40 FORMAT(//, ' BOND LENGTH BETWEEN ', 3,' OF A 1,F10.5, ' ANGSTROM',//)

CALL CHEMG(NTOT)

CALL MMDATA C WRITE FINAL CARTESIAN COORDINATE FOR CHEMGR 5 WRITE(6, ' (//lOX, ' 'FINAL CARTESIAN COORDINA

WRITE(6, ^• (4X, ' 'NO.' ',7X, ' 'ATOM' ' ,9X, ' *X' ', 1 9X, ' 'Y' ',9X, * 'Z' ',/)')

WRITE(6, * (I6,8X,A2,4X,3F10.5) ') 1 (I,ASYM(I), (COORD(J,I),J=l,3),I=l,NTOT) C

STOP

END C ROUTINE FOR OUTPUT IF INITIAL COORDINATES

SUBROUTINE IWRITE

PARAMETER (NT=150)

PARAMETER (NG=300)

CHARACTER*2 ASYM ,TITLE

INTEGER TYPEA,TYPEB ,TYP

COMMON/COOD/ COORD(3,NG) ,CHARGE(NG)

COMMON/TYPE/TYPEA(NT) ,TYPEB(NT) ,TYP(NG)

COMMON/SYMM/ASYM(NG) ,TITLE(40)

S ^U B ^ST ITUTESHEET

COMMON/CORD/XA(NT) ,YA(NT) ,ZA(NT) ,XB(NT) ,YB(N COMMON/FINAL/TXB(NT),TYB(NT) ,TZB(NT) ,CA(NT) , COMMON/INFO / NA,NB,IP1,IP2,IP3,IP4,R1,R2,CI WRITE(6, 15) TITLE 15 FORMAT(1H1,/////,

1 10X ' *************************************

2 10x ' COORDINATES OF SUPERMOLECULE

3 15X, 40A2 , /,

4 10X ' *************************************

C WRITE CARTESIAN COORDINATE

WRITE(6, ' (//10X, ' 'INITIAL CARTESIAN COORDI

WRITE(6, ' (4X, ' 'NO. ' ' ,7X, ' 'ATOM' ' ,9X, ^• 'X' ', 1 9X, ' ^• ¥' ',9X, ' 'Z' ',/)')

WRITE(6, ' (I6,8X,A2,4X,3F10.5) ') 1 (I,ASYM(I),XA(I),YA(I),ZA(I),I=1,NA) C

WRITE(6, ' (////10X, ' 'INITIAL CARTESIAN COOR

WRITE(6, ' (4X, ' 'NO. ' ',7X, ' 'ATOM' ',9X, ' 'X' ', 1 9X, ' 'Y' ',9X, ' 'Z' ',/)')

WRITE(6, ' (I6,8X,A2,4X,3F10.5) ' ) 1 (I,ASYM(NA+I), XB(I),YB(I),ZB(I),I=1,NB)

WRITE(6, ' (///)')

RETURN

END

SUBROUTINE OPTIM (X,N,ESCALE,IPRINT,ICON,E)

PARAMETER (NT=150)

DIMENSION W(1000),X(5) ,E(5)

MAXIT=100

DDMAG=0.1*ESCALE

SCER=0.05/ESCALE

JJ=N*N+N

JJJ=JJ+N

K=N+1

NFCC=1

IND=1

INN=1

DO 1 I=1,N

DO 2 J=1,N

W(K)=0.

IF(I-J)4,3,4

3 W(K)=ABS(E(I)) W(I)=ESCALE .

4 K=K+1

2 CONTINUE 1 CONTINUE

ITERC=1

ISGRAD=2

CALL CALCFX(N,X,F,EW,ECO,EIDS)

FKEEP=ABS(F)+ABS(F)

5 ITONE=l FP=F SUM=0. IXP=JJ

DO 6 1=1,N

SUBSTITUTE SHEET

IXP=IXP+1 W(IXP) =X(I) CONTINUE IDIRN=N+1 ILINE=1 DMAX=W(ILINE) DACC=DMAX*SCER DMAG=MIN (DDMAG,0.1*DMAX) DMAG=MAX(DMAG,20.*DACC) DDMAX=10.*DMAG

GO TO (70,70,71),ITONE DL=0.

D = DMAG

FPREV=F

IS=5

FA=F

DA=DL DD=D-DL DL=D K=IDIRN

DO 9 1=1,N

X(I)=X(I)+DD*W(K)

K=K+1 CONTINUE

CALL CALCFX(N,X,F,EW,ECO,EDIS)

NFCC=NFCC+1

-GO TO (10,11,12,13, 14,96),IS IF(F-FA)15,16,24 IF (ABS(D) -DMAX) 17,17,18 D=D+D GO TO 8 WRITE(6,19) FORMAT(5X,44HVA04A MAXIMUM CHANGE DOES NOT A GO TO 500 FB=F DB=D

GO TO 21 FB=FA DB=DA FA=F DA=D GO TO (83,23),ISGRAD D=DB+DB-DA

IS=1

GO TO 8 D=0.5*(DA+DB-(FA-FB)/(DA-DB))

IS=4

IF((DA-D)*(D-DB))25,8,8 IS=1

IF(ABS(D-DB)-DDMAX)8,8,26 D=DB+SIGN(DDMAX,DB-DA) IS=1

DDMAX=DDMAX+DDMAX DDMAG=DDMAG+DDMAG

SUBSTITUTE SHEET

IF(DDMAX-DMAX) 8, 8, 27 DDMAX=DMAX GO TO 8 IF(F-FA)28,23,23 FC=FB DC=DB FB=F DB=D

GO TO 30 IF(F-FB)28,28,31 FA=F DA=D

GO TO 30 IF(F-FB)32,10,10 FA=FB DA=DB

GO TO 29 DL=1.

DDMAX=5.

FA=FP

DA=-1.

FB=FHOLD

DB=0.

D=l. FC=F

DC=D A=(DB-DC)*(FA-FC) B= (DC-DA)*(FB-FC)

IF( (A+B)* (DA-DC)) 33, 33, 34 FA=FB DA=DB FB=FC DB=DC

GO TO 26 D=0.5*(A*(DB+DC) +B* (DA+DC) )/(A+B) DI=DB

FI=FB IF(FB-FC)44,44,43 DI=DC FI=FC GO TO (86,86,85),ITONE ITONE=2 GO TO 45 IF (ABS(D-DI)-DACC) 41,41,93 IF (ABS(D-DI)-0.03*ABS(D)) 41,41,45 IF ( (DA-DC) * (DC-D) ) 47,46,46 FA=FB DA=DB FB=FC DB=DC

GO TO 25 IS=2

IF ( (DB-D)*(D-DC)) 48,8,8 IS=3

SUBSTITUTE SHEET

GO TO 8 F=FI

D=DI-DL

DD=SQRT((DC-DB)*(DC-DA)*(DA-DB)/(A+B))

DO 49 1=1,N

X(I)=X(I)+D*W(IDIRN)

W(IDIRN)=DD*W(IDIRN)

IDIRN=IDIRN+1 CONTINUE

W(ILINE)=W(ILINE)/DD

ILINE=ILINE+1

IF(IPRINT-1)51,50,51 WRITE(6,52)ITERC,NFCC,F WRITE(7,52)ITERC,NFCC,F FORMAT (3X, ^• ITERATION' ,15,19, ' FUNCTION VAL WRITE(6,68) EW,ECO,EDIS WRITE(7,68) EW,ECO,EDIS FORMA (3X, 'REP. = ',F12.5,' ATT. =',F12.5, GO TO(51,53),IPRINT GO TO (55,38),ITONE IF (FPREV-F-SUM) 94,95,95 SUM=FPREV-F JIL=ILINE IF (IDIRN-JJ) 7,7,84 GO TO (92,72),IND FHOLD=F

IS=6

IXP=JJ

DO 59 1=1,

IXP=IXP+1

W(IXP)=X(I)-W(IXP) CONTINUE DD=1.

GO TO 58 GO TO (112,87),IND IF (FP-F) 37,91,91 D=2.*(FP+F-2.*FHOLD)/(FP-F)**2

IF (D*(FP-FHOLD-SUM)**2-SUM) 87,37,37 J=JIL*N+1

IF (J-JJ) 60,60,61 DO 62 I=J,JJ K=I-N W(K)=W(I) CONTINUE

DO 97 I=JIL,N W(I-1)=W(I) CONTINUE IDIRN=IDIRN-N ITONE=3 K=IDIRN IXP=JJ AAA=0.

DO 65 1=1,N IXP=IXP+1

SUBSTi ; _T \ i U _l T i E SHEET

w(K)=w(iχpy

IF (AAA-ABS(W(K)/E(I))) 66,67,67 AAA=ABS(W(K)/E(I)) K=K+1 CONTINUE DDMAG=1. W(N)=ESCALE/AAA ILINE=N GO TO 7 IXP=JJ AAA=0. F=FHOLD

DO 99 1=1,

IXP=IXP+1

X(I)=X(I)-W(IXP)

IF (AAA*ABS(E(I))-ABS(W(IXP))) 98,99,99 AAA=ABS(W(IXP)/E(I)) CONTINUE GO TO 72 AAA=AAA*(1.+DI)

GO TO (72,106),IND IF (IPRINT-2) 53,50,50 GO TO (109,88),IND IF (AAA-0.1) 89,89,76 GO TO (20, 116),ICON IND=2

GO TO (100, 101),INN INN=2

K=JJJ

DO 102 I=1,N

K=K+1

W(K)=X(I)

X(I)=X(I)+10.*E(I) CONTINUE

FKEEP=F

CALL CALCFX(N,X,F,EW,ECO,EDIS)

NFCC=NFCC+1

DDMAG=0.

GO TO 108 IF (F-FP) 35,78,78 WRITE(6,80) FORMAT (5X,37HVA04A ACCURACY LIMITED BY ERRO GO TO 500 IND=1 DDMAG=0.4*SQRT(FP-F)

IF(DDMAG.GE.1.0E60) DDMAG=1.0E60

ISGRAD=1 ITERC=ITERC+1

IF (ITERC-MAXIT) 5,5,81 WRITE(6,82)MAXIT FORMAT(15,30H ITERATIONS COMPLETED BY VA04A) IF (F-FKEEP) 500,500,110 F=FKEEP

DO 111 1=1,N

BSTi TE SHEET

JJJ=JJJ+1

X(I)=W(JJJ) 111 CONTINUE

GO TO 20 101 JIL=1

FP=FKEEP

IF (F-FKEEP) 105,78,104

104 JIL=2 FP=F F=FKEEP

105 IXP=JJ

DO 113 1=1,N

IXP=IXP+1

K=IXP+N

GO TO (114,115),JIL

114 W(IXP)=W(K) GO TO 113

115 W(IXP)=X(I) X(I)=W(K)

113 CONTINUE JIL=2 GO TO 92

106 IF (AAA-0.1) 20,20,107 20 WRITE(6,200)

WRITE(6,201)

WRITE(7,201) 201 FORMAT(5X, ' THE FUNCTION VALUE HAS BEEN MIN

WRITE(6,200) 200 FORMA (/1X, '******************************** 00 RETURN

107 INN=1

GO TO 35 END

SUBROUTINE PARM

PARAMETER (NT=150)

PARAMETER (NG=300)

CHARACTER*2 ASYM ,TITLE

INTEGER TYPEA,TYPEB ,TYP

COMMON/PARAM/ VEP(NG) ,VRA(NG)

COMMON/INFO / NA,NB,IP1,IP2,IP3,IP4,R1,R2,CI

COMMON/TYPE / TYPEA(NT),TYPEB(NT) ,TYP(NG)

DIMENSION VEPS(60) ,VRAD(60)

DATA VRAD/

1 1.900, 1.940, 1.940, 1.940, 1.500, 1.740, 1

2 -1.0 , 1.780, 1.740, 1.900, 0.930, 2.110, 1

3 1.820, 1.920, 1.200, 1.920, 1.325, 0.900, 1

4 1.740, 1.90 ,1.780 , 1.92, 1.82, -1.0 ,-1.0

5 -1.0 , -1.0 , -1.0 , -1.0 ,-1.0 , -1.0 , 1 DATA VEPS/

1 0.044, 0.044, 0.044, 0.044, 0.047, 0.050, 0

2 -1.0 , 0.066, 0.050, 0.030, 0.017, 0.202, 0

3 0.055, 0.044, 0.036, 0.044, 0.034, 0.015, 0

SUBSTITUTE SHEET

4 0.066, 0.044, 0.066, 0.044, 0.055,-1.0 ,-1.

5 -1.0 , -1.0 ,-1.0 , -1.0 ,-1.0 ,-1.0 , IF(JCON.GT.O) THEN

DO 10 J=l,JCON READ(5,*) ITYPE, EPS, RAD

VEPS(ITYPE) = EPS

VRAD(ITYPE) = RAD 10 CONTINUE

ENDIF

DO 20 1=1,NA

VEP(I) = VEPS(TYPEA(I) )

IF(VEP(I) .LE.0.0) THEN WRITE(7,25) TYPEA(I)

ENDIF

VRA(I) = VRAD(TYPEA(I))

IF(VRA(I) .LE.0.0) THEN WRITE(7,25) TYPEA(I) 25 FORMAT(4X, ' CHECK YOUR VAN DER WAAL DATA

ENDIF 20 CONTINUE C

DO 30 1=1,NB

VEP(NA+I) = VEPS(TYPEB(I))

IF(VEP(NA+I) .LE.0.0) THEN WRITE(7,25) TYPEB(I)

ENDIF

VRA(NA+I) = VRAD(TYPEB(I))

IF(VRA(NA+I) .LE.0.0) THEN WRITE(7,25) TYPEB(I)

ENDIF 30 CONTINUE

RETURN

END

SUBROUTINE ENERGY(ETOT,EV,ETOTl,EDIS) C FUNCTION PROGRAM FOR SUPER-MOLECULE

PARAMETER (NT=150)

PARAMETER (NG=300)

CHARACTER*2 ASYM ,TITLE

INTEGER TYPEA,TYPEB ,TYP

COMMON/COOD/ COORD(3,NG) ,CHARGE(NG) _Λ

COMMON/TYPE/TYPEA(NT) ,TYPEB(NT) ,TYP(NG)

COMMON/CORD/XA(NT) ,YA(NT) ,ZA(NT) ,XB(NT) ,YB(N

COMMON/FINAL/TXB(NT) ,TYB(NT) ,TZB(NT) ,CA(NT) ,

COMMON/PARAM/ VEP(NG) ,VRA(NG)

COMMON/INFO / NA,NB,IPl,IP2,IP3,IP4,R1,R2,CI

COMMON/PATH/MYWAY C CALCULATION OF VAN DER WALLS ENERGY(ONLY 1-5

DIELC=78.5 C GO TO (1 ,2), MYWAY

X1=XA(IP2)

Y1=YA(IP2)

Z1=ZA(IP2)

X2=TXB(IP4)

Y2=TYB(IP4)

SUBSTITUTE SHEET

Z2=TZB(IP4)

ET0T1=DIST(XI,Yl,Zl,X2,Y2,Z2) ET0T1= ABS(ET0T1-R2) *500.0 IF(MYWAY.EQ.l) THEN

ET0T=ET0T1

RETURN ENDIF

2 EV=0.0

EC=0.0

DO 500 1=1,NA

XI=XA(I)

YI=YA(I)

ZI=ZA(I)

DO 500 K=1,NB

XK=TXB(K)

YK=TYB(K)

ZK=TZB(K)

RIK=DIST(XI,YI,ZI,XK,YK,ZK)

ECOUL=332.0*CHARGE(I)*CHARGE(NA+K)/(DIELC*RI

VEPI=VEP(I)

VEPK=VEP(NA+K)

VRAI=VRA(I)

VRAK=VRA(NA+K)

EPS=SQRT(VEPI*VEPK)

RV=VRAI+VRAK

P=RV/RIK

IF(P.GT.3.31) GO TO 30

IF(P.LT.0.072) THEN E=EPS*(-2.25*P**6) GO TO 35

ENDIF

E=EPS*(290000.0*EXP(-12.5/P) -2.25*P**6)

GO TO 40 30 E=EPS*336.176*P*P 35 CONTINUE 40 EV = EV+E

EC = EC +ECOUL 500 CONTINUE

ETOT=EV + ETOT1 + EC C ETOT=EV + ETOT1

X1=XA(IP2)

Y1=YA(IP2)

Z1=ZA(IP2)

X2=TXB(IP4)

Y2=TYB(IP4)

Z2=TZB(IP4)

EDIS=DIST (XI , Yl , Z 1 , 2 , Y2 , Z2 )

RETURN

END C FUNCTION DIST

FUNCTION ^" DIST (XI , Yl , Z 1 , X2 , Y2 , Z2 )

X=X1 -X2

Y=Y1-Y2

SUBSTITUTE SHEET

Z=Z1-Z2

DIST=SQRT(X*X+Y*Y+Z*Z)

RETURN

END

SUBROUTINE CALCFX(NVAR,X,ETOT,EV,EC,EDIS)

PARAMETER (NT=150)

PARAMETER (NG=300)

COMMON/CORD/XA(NT) ,YA(NT) ,ZA(NT) ,XB(NT) ,YB(N

COMMON/FINAL/TXB(NT) ,TYB(NT) ,TZB(NT) ,CA(NT) ,

COMMON/INFO / NA,NB,IPl,IP2,IP3,IP4,Rl,R2 Cl

DIMENSION X(5) ,TX(150) ,T (150) ,TZ(150) ,XROT(

DIMENSION CTX(150) ,CTY(150) ,CTZ (150) C CONVERSION OF POLAR COORDINATE TO CARTECIAN COOR

DX=Rl*SIN(X(l))*COS(X(2))

DY=R1*SIN(X(1))*SIN(X(2))

DX=Rl*COS(X(l)) C FIXING OF PROBE 2 APART FROM PROBE 1 BY Rl ANG.

PX=XA(IP1)+DX

PY=YA(IP1)+DY

PZ=ZA(IP1)+DZ C CALCULATE DISTANCE VECTORS BETWEEN PROBE P(PX,P

DVX=PX-XB(IP3)

DVY=PY-YB(IP3)

DVZ=PZ-ZB(IP3) C PARALLEL MOVEMENT OF B MOLECULE BY (DVX,DVY,DVZ

DO 10 IM=1,NB

TX(IM) = XB(IM) + DVX

TY(IM) = YB(IM) + DVY

TZ(IM) = ZB(IM) + DVZ 10 CONTINUE C MOVE TO MAKE AN ORIGIN OF PROBE3(IP3) IN

DO 20 10=1,NB

IF(IO.EQ.IP3) GO TO 20

TX(IO)=TX(IO)-TX(IP3)

TY(IO)=TY(IO)-TY(IP3)

TZ(IO)=TZ(IO)-TZ(IP3) 20 CONTINUE

TX(IP3) = 0.0D0

TY(IP3) = 0.0D0

TZ(IP3) = 0.0D0 C ROTATION

CSX= COS(X(3))

SSX= SIN(X(3) )

CSY= COS(X(4))

SSY= SIN(X(4))

CSZ= COS(X(5))

SSZ= SIN(X(5)) C X ROTATION

DO 30 1=1,9 30 . XROT(I) = 0.0

XROT(l) = 1.0

XROT(5) = CSX

XROT(6) =-SSX

XROT(8) = SSX

SUBSTITUTE SHEET

XR0T(9) = CSX DO 40 1=1,9 0 YROT(I) = 0.0 YROT(l) = CSY YROT(3) = SSY YROT(5)= 1.0 YROT(7)= -SSY YROT(9) = CSY DO 50 1=1,9

50 ZROT(I)= 0.0 ZROT(l) = CSZ ZROT(2) = SSZ ZROT(4) =-SSZ ZROT(5) = CSZ ZROT(9) = 1.0 DO 60 J=1,NB COXX = XROT(l)*TX (J ) +XROT(2)* TY(J ) +X COXY = XROT(4)*TX (J ) +XROT(5)* TY(J ) +X COXZ = XROT(7)*TX (J ) +XROT(8)* TY(J ) +X COYX. = YROT(1)*COXX +YROT(2)*COXY +YROT(3)*C COYY = YROT(4)*COXX +YROT(5)*COXY +YROT(6)*C COYZ = YROT(7)*COXX +YROT(8)*COXY +YROT(9)*C CTX(J)=ZROT(1)*COYX +ZROT(2)*COYY +ZROT(3)*C CTY(J)=ZROT(4)*COYX +ZROT(5)*COYY +ZROT(6)*C CTZ( )=ZROT(7)*COYX +ZROT(8)*COYY +ZROT(9)*C

60 CONTINUE C RETURN TO POINT P

DO 70 1= 1,NB TXB(I) = CTX(I) PX TYB(I) = CTY(I) PY TZB(I) = CTZ(I) PZ

70 CONTINUE

C

CALL ENERGY(ETOT,EV,EC,EDIS)

RETURN

END

SUBROUTINE COUL(ETOT,ER,EA)

PARAMETER (NT=150)

PARAMETER (NG=300)

CHARACTER*2 ASYM ,TITLE

COMMON/CORD/XA(NT) ,YA(NT) ,ZA(NT) ,XB(NT) ,YB(N

COMMON/FINAL/TXB(NT) ,TYB(NT) ,TZB(NT) ,CA(NT) ,

COMMON/INFO / NA,NB,IP1,IP2,IP3,IP4,R1,R2,CI

ER=0.0

DO 10 1=1,NA

X1=XA(I)

Y1=YA(I)

Z1=ZA(I)

DO 20 J=1,NB

X2=TXB(I)

Y2=TYB(I)

Z2=TZB(I)

ER =ER +(CA(I)*CB(J))/DIST(X1,Y1,Z1,X2,Y2,Z2

SUBSTITUTE

20 CONTINUE 10 CONTINUE

EA =0.0

X1=XA(IP2)

Y1=YA(IP2)

Z1=ZA(IP2)

X2=TXB(IP4)

Y2=TYB(IP4)

Z2=TZB(IP4)

EA =EA +(CIP2*CIP4)/DIST(X1,Y1,Z1,X2,Y2,Z2)

ETOT=ER + EA

RETURN

END

: SUBPROGRAM TO GENERATE ATOM TYPE AND NET ATO SUBROUTINE CHARG PARAMETER (NG=300) PARAMETER (NT=150) INTEGER TYPEA,TYPEB ,TYP COMMON/COOD / COORD(3,NG) ,CHARGE(NG) COMMON/TYPE / TYPEA(NT) ,TYPEB(NT),TYP(NG) COMMON/INFO / NA,NB,IPl,IP2,IP3,IP4,R1,R2,CI DIMENSION DCHB(35) DATA DCHB/ 0.241, 0.0 , 0.515, 0.0 , 0.0

1 -0.267,-0.509, 0.0 ,-0.692 -0 .5

2 -0.135, 0.0 , 0.0 , 0.0 -0 . 6

3 0.0 , 0.243, 0.15 , 0.131 -0. 5

4 0.0 ,-0.622, 0.515, 0.0 0 .0 DO 20 1=1,NA

CHARGE(I) =DCHB(TYPEA(I) ) 20 CONTINUE C SPECIAL SIDE CHAIN FOR CARBONE C CHARGES FOR HYDANTON

DO 30 1=1,NB

CHARGE(NA+I) =DCHB(TYPEB(I) ) 30 CONTINUE C CHARGE(NA+1) = -0.36 C CHARGE(NA+2) = 0.44 C CHARGE(NA+3) = -0.41 C CHARGE(NA+4) = 0.58 C CHARGE(NA+5) = -0.31 C CHARGE(NA+6) = 0.03 C CHARGE(NA+7) = -0.41 C CHARGE(NA+20)= 0.19 C CHARGE(NA+21)= 0.20 40 CHARGE(IP2)= CIP2

CHARGE(NA+IP4) = CIP4 WRITE(6,22) (CHARGE(I),1=1, 12) 22 FORMAT(3X,6F10.4)

RETURN

END C

SUBROUTINE SYMBOL

PARAMETER (NT=150)

SUBSTITUTE SHEET

PARAMETER (NG=300)

CHARACTER*2 ASYM ,TITLE

INTEGER TYPEA,TYPEB,TYP,HH,NN,00,CC

COMMON/COOD/ COORD(3,NG) ,CHARGE(NG)

COMMON/TYPE/TYPEA(NT) ,TYPEB(NT),TYP(NG)

COMMON/SYMM/ASYM(NG) ,TITLE(40)

COMMON/INFO / NA,NB,IPl,IP2,IP3,IP4,R1,R2,CI

DIMENSION HH(6) ,NN(7) ,00(6) ,CC(10)

DATA HH/5, 14,23,21,24,25/

DATA NN/8,9,13,19,26,32,28/

DATA 00/6,7,11,12,28,30/

DATA CC/1,2,3,4, 16,20,22,27,29,31/

NT0T=NA+NB

DO 1 1=1,NA

TYP(I)=TYPEA(I) CONTINUE

DO 2 1=1,NB

TYP(NA+I)=TYPEB(I) CONTINUE 1=0 . CONTINUE 1=1+1

IF(I.GT.NTOT) GO TO 9 DO 20 Kl=l,6 IF(TYP(I) .EQ.HH(Kl)) THEN ASYM(I)=' H' GO TO 10 ENDIF CONTINUE

DO 30 Kl=l,7 IF(TYP(I) .EQ.NN(Kl)) THEN ASYM(I)=' N' GO TO 10 ENDIF CONTINUE

DO 40 Kl=l,6 IF(TYP(I) .EQ.OO(Kl)) THEN ASYM(I)=' 0' GO TO 10 ENDIF CONTINUE

DO 50 Kl=l,10 IF(TYP(I) .EQ.CC(Kl)) THEN ASYM(I)=' C GO TO 10 ENDIF CONTINUE

IF(TYP(I) .EQ.15) THEN ASYM(I)= ^• S' GO TO 10 ELSE

WRITE(6,100) TYP(I) ,1 FORMAT(3X, ' UNDEFINED ATOM TYPE. : ',13,' 0 ENDIF

SUBSTITUTE SHEET

CONTINUE RETURN END SUBROUTINE CHEMG(NTOT)

CHARACTER NAME1*2, NAME2*3 ,CTEMP*80 ,TEMP*

PARAMETER (NG=300)

PARAMETER (NT=150)

INTEGER IH,IN,IC,10,IS,HH,NN,CC,00,SS

COMMON/COOD/ COORD(3,NG) ,CHARGE( G)

COMMON/SYMM/ASYM(NG) ,TITLE(40)

J=0

I0NE=1

WRITE(10,1) NTOT,IONE,ΪONE

FORMAT(//,I3,I5,/,I6)

DO 33 I = l,NTOT

READ(4, ' (1X,3F10.4,5X,A1) ' ,END=999) X,Y,Z,NA X=COORD(l,I) Y=COORD(2,I) Z=COORD(3,I) NAME1=ASYM(I)

IF( NAME1.EQ.' F') GO TO 33 J=J+1

IF (NAME1.EQ.' H') THEN IH = IH + 1

WRITE(NAME2, ' (13) ') IH ELSEIF (NAME1.EQ.' N') THEN IN = IN + 1

WRITE(NAME2, ' (13) ') IN ELSEIF (NAME1.EQ.' C) THEN IC = IC + 1

WRITE(NAME2, ' (13) ' ) IC ELSEIF (NAME1.EQ.' 0') THEN 10 = 10 + 1

WRITE(NAME2, ' (13) ') 10 ELSEIF (NAME1.EQ.' S') THEN IS = IS + 1

WRITE(NAME2, ' (13) ' ) IS ELSE

WRITE(6, ' (' 'You have a problem on line' ',14, &')I ENDIF

IF (NAME2(1:1) .EQ. ' ') THEN NAME2(1:1) = NAME2(2:2) NAME2(2:2) = NAME2(3:3) NAME2(3:3) = ' ' IF (NAME2(1:1) .EQ. ' ') THEN NAME2(1:1) = NAME2(2:2) NAME2(2:2) = ' ' ENDIF ENDIF WRITE(10, ' (I4,A2,A3,1X,3F10.4) ') J,NAME1,NAM

S ^UB STITUTESHEET

33 CONTINUE C ENDIF RETURN END SUBROUTINE WRIT(X,Y,Z) DIMENSION X(150) ,Y(150) ,Z(150) DO 5 1=1,6 5 WRITE(6,10) X(I),Y(I),Z(I) 10 FORMAT(4X,3F12.5) RETURN END C

SUBROUTINE COMBIN (NA,NB) DIMENSION ICON(16) ,IAT1(150) ,IAT2(150) CHARACTER*2 TT(30) READ(4,10) NA 10 FORMAT(62X,13) READ(5,10) NB NTOT=NA+NB IONE=l IFOUR =4 TIME=100.0

WRITE(16,20) NTOT,IFOUR,IONE,TIME 20 FORMAT(60X,I5,I2,3X,I5,F5.0)

READ(4,30) NCONA,NATA,NSPA 30 FORMAT(I5,20X,I5,15X,I5)

READ(5,30) NCONB,NATB,NSPB

NCOT=NCONA+NCONB

NATT=NATA+NATB

NSPT=NSPA+NSPB

WRITE(16,30) NCOT,NATT,NSPT

IF(NSPA.NE.O) THEN

DO 50 I=1,NSPA ^" READ(4,40) TT 40 FORMAT(30A2)

WRITE(16,40) TT 50 CONTINUE ENDIF

IF(NSPB.NE.O) THEN DO 60 I=1,NSPB READ(5,40) TT WRITE(16,40) TT 60 CONTINUE ENDIF

DO 70 IA=l,NCONA READ(4,75) (ICON(I),1=1,16) 75 FORMAT(1615) DO 80 IZ=1,16 ISZ=16

IF( ICON(IZ) .EQ.0) THEN ISZ=IZ -1 GO TO 85 ENDIF 80 CONTINUE

SUBSTITUTE SHEET

85 WRITE(16,75) (ICON(I) ,1=1,ISZ) 70 CONTINUE

DO 90 IB=l,NCONB READ(5,75) (ICON(I) ,1=1,16) DO 100 IZ=1,16 ISZ=16

IF( ICON(IZ) .EQ.0) THEN ISZ=IZ -1 GO TO 95 ENDIF 100 CONTINUE 95 DO 110 1=1,ISZ 110 ICON(I)= ICON(I)+NA

WRITE(16,75) (ICON(I) ,1=1,ISZ) 90 CONTINUE

READ(4,75) (IAT1(I) ,IAT2(I) ,I=1,NATA) NATA1=NATA+1

READ(5,75) (IAT1(I) ,IAT2(I) ,I=NATA1,NATT) DO 120 IL=NATA1,NATT IATl(IL) = IAT1(IL)+NA IAT2(IL) = IAT2(IL)+NA 120 CONTINUE

WRITE(16,75) (IAT1(I) ,IAT2(I),1=1,NATT) RETURN END C SUBROUTINE FOR MM INPUT SUBROUTINE MMDATA PARAMETER (NT=150) PARAMETER (NG=300) CHARACTER*2 ASYM ,TITLE INTEGER TYPEA,TYPEB ,TYP COMMON/COOD / COORD(3,NG) ,CHARGE(NG) COMMON/TYPE / TYPEA(NT) ,TYPEB(NT) ,TYP(NG) COMMON/SYMM / ASYM(NG) ,TITLE(40) COMMON/PARAM/ VEP(NG) ,VRA(NG)

COMMON/CORD / XA(NT) ,YA(NT) ,ZA(NT) ,XB(NT) ,YB COMMON/FINAL/ TXB(NT) ,TYB(NT),TZB(NT),CA(NT) COMMON/INFO / NA,NB,IPl,IP2,IP3,IP4,R1,R2,CI NTOT=NA+NB

WRITE(16,20) (COORD(1,1),COORD(2,I) ,COORD(3, 20 FORMAT(2(3F10.5,I5,5X)) RETURN END

SUBSTITUTE SHEbT