JP2007272627A - Screening method and screening system - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To efficiently select a bonding form that is a candidate from a plurality of bonding forms generated by docking simulation. <P>SOLUTION: This screening system 100 includes a structure information reception part 101, a bonding form generation processing part 103, a parameter calculation part 111, a selection part 113, and an output part 115. The bonding form generation processing part 113 performs short-time structure optimization to the bonding format of a compound and a protein generated by the docking simulation. The parameter calculation part 111 measures a filling degree of the compound before and after the optimization to the protein, and selects the bonding form wherein a structure is optimized such that filling is further performed. <P>COPYRIGHT: (C)2008,JPO&INPIT

Description

  The present invention relates to a screening method and a screening system, and more particularly to in silico screening of low molecular compounds.

  A method for predicting the binding ability of a protein and a compound using a computer when the structure of the binding site of the protein is clear is called SBDD (structure based drug design). There are various methods of SBDD, each having different accuracy and different processing speed.

  Searching for drug candidate compounds from a low molecular compound library using SBDD is called in silico screening.

  In the initial stage of in silico screening, a method called docking simulation in which a protein binding site is bound to a low molecular weight compound is used.

  Docking simulation is positioned as a technique capable of high-speed processing, although the accuracy is relatively rough among various SBDD techniques. In addition, various binding modes can be predicted based on the shapes of proteins and low-molecular compounds by docking simulation.

Further, as a method for improving the screening accuracy in the docking simulation, there is a method described in Patent Document 1. Patent Document 1 describes scoring by docking simulation after obtaining the three-dimensional structure and ligand structure of a target protein. And when docking simulation does not output a stable structure as a complex of target protein, after obtaining scoring, it is said that optimization of each complex structure in solution is performed by molecular dynamics simulation .
JP-A-2005-181104

  When using docking simulation, it is conceivable to predict and generate as many coupling modes as possible and determine which coupling mode is appropriate using a score function or the like. However, since the accuracy of the score function itself is relatively rough, in the end, many coupling modes are shifted to a stage where more detailed calculation is performed. As a result, it was difficult to perform in silico screening with a certain degree of accuracy for a large number of small molecule compounds of 1 million or more in a realistic time.

  In addition, if the optimization calculation is performed for the enormous coupling mode generated by the docking simulation as in the technique described in Patent Document 1, it takes a long time for the optimization calculation. For this reason, it is not realistic as a method for efficiently narrowing down the coupling mode within a limited time.

  The present invention has been made in view of the above circumstances, and an appropriate binding mode candidate is efficiently selected from a plurality of binding modes with low molecular weight compounds for proteins.

According to the present invention,
Generating a plurality of binding modes of the low molecular weight compound to the protein by docking simulation of the protein and the low molecular weight compound;
Calculating a parameter relating to the binding mode for the plurality of generated binding modes;
Starting a structure optimization process for the plurality of generated coupling modes;
Stopping the structure optimization process before the structure optimization process is completed;
Calculating the parameters for the plurality of coupling modes after stopping the structure optimization process;
For each coupling mode, comparing the parameters before and after the start of the structure optimization process, and selecting a coupling mode in which a predetermined change has occurred in the parameter from the plurality of coupling modes;
A screening method is provided.

Moreover, according to the present invention,
A binding state generation processing unit that obtains structural information of proteins and low-molecular compounds, and generates and processes binding modes between the proteins and the low-molecular compounds;
A parameter calculation unit for calculating a parameter related to the coupling mode for the plurality of coupling modes;
Based on the parameters calculated by the parameter calculation unit, a selection unit that selects a coupling mode from the plurality of coupling modes;
Including
The combined state generation processing unit
A binding mode generation unit that generates a plurality of binding modes of the low molecular weight compound to the protein by docking simulation of the protein and the low molecular weight compound;
An optimization processing unit that performs structure optimization processing of the coupling mode generated by the coupling mode generation unit;
An optimization processing unit control unit for stopping the optimization processing in the optimization processing unit at a predetermined stage;
Including
The selection unit acquires the parameters before and after the start of the structure optimization process for each coupling mode, compares the parameters before and after the start of the structure optimization process, A screening system for selecting a binding mode in which a predetermined change has occurred from the plurality of binding modes is provided.

  In the present invention, a plurality of binding modes of low-molecular compounds to proteins are generated by docking simulation. For this reason, it is possible to generate a large number of candidate binding modes.

  Furthermore, in the present invention, after generating the coupling mode, the optimization processing unit is started and stopped at a predetermined stage. By stopping the optimization process that takes a long time until completion, it is possible to reduce the time required for selecting the coupling mode. Then, parameters related to the coupling mode are calculated before and after the optimization process, and the parameters before and after the optimization process are compared for each coupling mode, and the coupling mode in which a predetermined change has occurred in the parameter is selected. . Since the change in the parameter corresponding to the change in the coupling mode caused by the short optimization process is used for the selection of the coupling mode, the candidates for the coupling mode can be efficiently narrowed down without performing the optimization process completely. .

  As described above, according to the present invention, it is possible to narrow down efficiently while ensuring the accuracy of narrowing down the coupling mode.

  As described above, according to the present invention, candidate combining modes can be efficiently selected from a plurality of combining modes generated by docking simulation.

  Embodiments of the present invention will be described below with reference to the drawings. In all the drawings, common constituent elements are denoted by the same reference numerals, and description thereof is omitted as appropriate.

FIG. 1 is a functional block diagram showing the configuration of the screening system in the present embodiment.
The screening system 100 shown in FIG. 1 includes a structure information receiving unit 101, a coupling style generation processing unit 103, a parameter calculation unit 111, a selection unit 113, and an output unit 115.
The structure information receiving unit 101 acquires information on the three-dimensional structure of the protein and information on the structure of the low molecular compound.
The structure information receiving unit 101 may receive an external input for information regarding these structures, or may acquire data stored in the storage unit as will be described later with reference to FIG.

  The binding mode generation processing unit 103 acquires the structural information of the protein and the low molecular compound received by the structure information receiving unit 101, and generates and processes the binding mode between the protein and the low molecular compound. The combination format generation processing unit 103 includes a combination format generation unit 105, an optimization processing unit 107, and an optimization processing unit control unit 109.

  The binding mode generation unit 105 acquires the structural information of the protein and the low molecular compound received by the structural information reception unit 101, and generates a plurality of binding modes of the low molecular compound to the protein by docking simulation of the protein and the low molecular compound. To do.

  The optimization processing unit 107 performs structure optimization processing of the coupling mode generated by the coupling mode generation unit 105.

  Further, the optimization processing unit control unit 109 stops the optimization processing in the optimization processing unit 107 at a predetermined stage.

  In the binding mode generation processing unit 103, for example, the binding mode generation unit 105 generates a plurality of binding modes by general software that performs docking simulation of a protein and a low molecular compound, and the optimization processing unit 107 Short-term structural optimization that can be realized with molecular dynamics software.

The parameter calculation unit 111 calculates a parameter related to the coupling mode for a plurality of coupling modes.
First, the parameter calculation unit 111 acquires data relating to the coupling mode generated by the coupling mode generation unit 105 and calculates the parameters. This is a pre-optimization parameter obtained for the state before the optimization processing unit 107 performs the optimization processing.

  In addition, the parameter calculation unit 111 calculates the parameters for each coupling mode generated by the coupling mode generation unit 105. In addition, the parameter calculation unit 111 calculates a parameter in a state after the structure optimization process is performed in the optimization processing unit 107 for each coupling mode, and sets it as a post-optimization parameter.

The parameter calculation unit 111 calculates, for example, a parameter related to the degree of packing of the low molecular compound in the binding mode with respect to the protein.
Here, the degree of filling is the degree of presence of protein atoms around the ligand, and the parameter relating to the degree of filling is a parameter that reflects the number of protein atoms existing in a certain distance from the ligand.

  The parameter relating to the filling degree may be a parameter indicating the filling degree itself or a parameter having a correlation with the filling degree. Examples of the parameter having a correlation with the degree of filling include the hydrophobic interaction energy Ehp in Japanese Patent Application No. 2006-038922. The hydrophobic interaction energy Ehp is expressed as a function of the degree of filling as shown below, and has a strong correlation with the filling rate. Hereinafter, a case where the parameter relating to the filling degree is Ehp will be described as an example.

The hydrophobic interaction energy Ehp is represented by the following formula (1).
Ehp = −εhpΣiHiS (μi) (1)
In the above formula (1), Σi represents the sum related to the i-th ligand atom. Further, εhp is an energy parameter, H is a hydrophobic parameter specific to the atom type, S is a degree of burying, and μ is a degree of filling. By making Ehp a smooth function of distance, the gradient can be easily calculated.

The degree of burying S indicates the abundance ratio of atoms around the focused atom, and the larger the value of S, the more the protein pocket region is filled.
The burial degree S is expressed by the following formulas (2) to (4) as a function of the filling degree μ.
S (μ) = 0 (0 ≦ μ ≦ μ1) (2)
S (μ) = (1/2) {1 + cosπ ((μ2−μ) / (μ2−μ1))} (μ1 ≦ μ ≦ μ2) (3)
S (μ) = 1 (μ ≧ μ2) (4)

  From the above formulas (2) to (4), S is a function that smoothly connects between μ1 and μ2. S can be regarded as the weight of the interaction energy parameter.

The filling degree μ is expressed by the following formula (5).
μi = Σjρ (rij) (5)

In the above equation (5), μ is a function of ρ. ρ is a function of the distance for determining whether the i-th ligand atom is at the interaction distance with the receptor atom and the j-th atom representing a ligand atom other than the i-th, and the following equations (6) to (8) Indicated by
ρ (rij) = 1 (0 ≦ rij ≦ σ1) (6)
ρ (rij) = (1/2) {1 + cosπ ((rij−σ1) / (σ2−σ1))} (σ1 ≦ rij ≦ σ2) (7)
ρ (rij) = 0 (rij ≧ σ2) (8)

  From the above formulas (6) to (8), ρ is a function that smoothly connects between σ1 and σ2. σ1 and σ2 are parameters specific to the atom pair.

  The parameter calculation unit 111 measures the change in the filling degree before and after the structure optimization by using, for example, software for calculating the interaction energy between the protein and the compound or general software for measuring the filling rate of the compound with respect to the protein.

  Returning to the description of the screening system 100, the selection unit 113 selects a coupling mode from a plurality of coupling modes based on the parameters calculated by the parameter calculation unit 111. For example, the selection unit 113 selects a coupling mode from among a large number of coupling modes, and narrows down the coupling mode candidates to the selected one.

  The selection unit 113 acquires parameters before and after the start of the structure optimization process for each coupling mode. That is, a pair of pre-optimization parameter value and post-optimization parameter value corresponding to each combination mode is acquired. Then, the parameters before and after the start of the structure optimization process are compared, and a coupling mode in which a predetermined change occurs in the parameter is selected from a plurality of coupling modes.

  The output unit 115 outputs information related to the coupling mode selected by the selection unit 113 to the user.

  In addition, the screening system 100 may include a display unit that displays information related to the coupling mode selected by the selection unit 113.

  The screening system 100 may further include a storage unit. FIG. 2 is a diagram illustrating an example of a configuration of a screening system including a storage unit. The basic configuration of FIG. 2 is the same as that of FIG. 1 except that a storage unit 117 including a structure information storage unit 119, a coupling format storage unit 121, and a parameter storage unit 123 is further provided.

  The structure information storage unit 119 stores information on the three-dimensional structure of the protein and information on the structure of the low molecular compound.

  The binding mode storage unit 121 stores information on the binding mode between the protein generated by the binding mode generation processing unit 103 and the low molecular weight compound. The combination format storage unit 121 also stores information related to the combination format after being optimized by the optimization processing unit 107. Further, the coupling mode storage unit 121 may store data related to a coupling mode calculation method and an optimization processing method.

  The parameter storage unit 123 stores parameter values related to the filling degree calculated by the parameter calculation unit 111. The parameter storage unit 123 stores the parameter calculation results before and after the optimization process. The parameter storage unit 123 may store parameter calculation formula data.

  Next, an in silico screening method using the screening system of the present embodiment will be described with reference to FIG. 3 and FIG. Hereinafter, the case of the screening system of FIG. 2 will be described as an example. FIG. 3 is a flowchart showing the procedure of the screening method in the present embodiment. FIG. 4 is a flowchart for explaining step 104 in FIG. 3 in more detail.

The screening method of this embodiment includes the following procedures:
S101: Acquisition of structural information of proteins and low-molecular compounds,
S102: Generation of a binding mode between a protein and a low molecular weight compound,
S103: Calculation of parameters related to the binding mode for the generated binding mode,
S104: Optimizing process of coupling mode,
S105: Calculation of parameters related to the coupling mode for the coupling mode after the optimization process, and S106: Selection of a coupling mode from a plurality of coupling modes.

The optimization process in step 104 is:
Each step of S111: Start of optimization processing and S112: Stop of optimization processing are included.

  In step 101, the structure information receiving unit 101 receives information related to the three-dimensional structure of the protein and information related to the structure of the low molecular compound. At this time, the structure information receiving unit 101 may receive an external input, or may refer to the structure information storage unit 119. Further, the optimization processing unit control unit 109 may store data related to the structure input from the outside in the structure information storage unit 119.

  In step 102, the binding mode generation processing unit 103 acquires the structural information received by the structural information receiving unit 101, and generates a plurality of binding modes of the low molecular compound to the protein by docking simulation of the protein and the low molecular compound. . The combination format generation processing unit 103 may acquire the structure information directly from the structure information receiving unit 101 or may refer to the structure information storage unit 119. Further, the coupling mode generation processing unit 103 may store data relating to the generated coupling mode in the coupling mode storage unit 121.

  In step 103, the parameter calculation unit 111 acquires information regarding a plurality of coupling modes generated by the coupling mode generation processing unit 103, and calculates parameters regarding the coupling mode for the acquired coupling modes.

  The parameter calculation unit 111 calculates, for example, a parameter related to the degree of filling of the low molecular compound with respect to the protein in each binding mode. More specifically, the parameter relating to the degree of filling is the hydrophobic interaction energy between the above-described protein and the low molecular weight compound.

  Among the optimization processes of step 104, in step 111, the optimization processing unit 107 acquires information on a plurality of coupling modes generated by the coupling mode generation processing unit 103, and for each of the acquired plurality of coupling modes, Start the structure optimization process. The optimization processing unit 107 may store the data of the coupling mode after the structure optimization process in the coupling mode storage unit 121.

  In step 112, the optimization processing unit control unit 109 stops the structural optimization processing in the optimization processing unit 107 before the structural optimization processing is completed. For example, the optimization processing unit control unit 109 stops the structure optimization processing after a predetermined time has elapsed from the start of the optimization processing. Further, the optimization processing unit control unit 109 may stop the processing when the structure optimization processing has advanced to a predetermined stage.

  In step 105, the optimization processing unit control unit 109 acquires data of a plurality of coupling modes after the structure optimization process is stopped, and calculates parameters after optimization for each coupling mode. The parameter calculation unit 111 may store the calculated parameter in the parameter storage unit 123.

  In step 106, the selection unit 113 refers to the parameter storage unit 123, acquires parameter values before and after the start of the structure optimization process for each coupling mode, compares them, and compares the parameters. A coupling mode in which a predetermined change occurs is selected from a plurality of coupling modes. Further, the selection unit 113 may send the data of the selected coupling mode to the output unit 115.

  In step 106, the selection unit 113 selects a binding mode in which the parameters related to the degree of filling are changed so that the low molecular weight compound is filled in the protein after the structure optimization process is stopped and before the start of the structure optimization. To do. More specifically, the selection unit 113 has a binding mode in which the value of the hydrophobic interaction energy after the structure optimization process is stopped is smaller than the value of the hydrophobic interaction energy before the structure optimization process is started. select. By the above procedure, an appropriate binding mode can be selected from a plurality of binding modes generated by the screening system docking simulation.

A flow chart summarizing the above procedure is shown in FIG.
As shown in FIG. 5, in this embodiment, a plurality of coupling modes are generated by docking simulation (S121), structure optimization calculation is performed for all coupling modes (S122), and all coupling modes are performed. In contrast, the filling rate before and after the structure optimization is calculated (S123). For bonding modes with improved filling rate due to structural optimization (Yes in S124), they are adopted as appropriate bonding modes (S125), and for bonding modes without improved packing rate (No in S124) Not adopted (S126).

  As described above, in this embodiment, as a determination method for narrowing down the binding mode candidates between the protein and the compound, the structure optimization in a short time is performed on the binding mode generated by the docking simulation. In the determination, a parameter indicating the degree of filling is used as an index, and the degree of filling with respect to the protein of each compound of the binding mode changed by the structure optimization in a short time is compared. It is possible to select an appropriate bonding mode by performing a short-term structural optimization on the bonding mode generated by the docking simulation and measuring the filling degree before and after that.

  As mentioned above, although embodiment of this invention was described with reference to drawings, these are the illustrations of this invention, Various structures other than the above are also employable.

  For example, in the above description, the case where the hydrophobic interaction energy shown in the above formula (1) is used as the parameter indicating the degree of filling of the ligand has been described as an example. However, the parameter indicating the degree of filling of the ligand is not limited thereto. Alternatively, other parameters that correlate with the degree of filling may be used. As another parameter, for example, a filling factor of a ligand can be used. The ligand filling rate is the volume of the region in which the ligand is embedded relative to the volume of the pocket region formed by the protein.

  Further, in the screening system 100, the connection mode generation processing unit 103 may further include a clustering unit that clusters and classifies the plurality of connection modes generated by the connection mode generation unit 105. The clustering unit may perform clustering so that similar connection modes are grouped together. Further, the optimization processing unit 107 may perform optimization processing on the representative of each cluster classified by the clustering unit. Further, the parameter calculation unit 111 may calculate parameters for the representatives of the clusters classified by the clustering unit. In this way, it is possible to efficiently narrow down a large number of coupling modes generated by the docking simulation in a shorter time.

  A docking simulation was performed on a protein registered in PDB (Protein Data Bank) and a ligand known to bind to the binding site of the protein, and 50 binding modes were generated.

  Clustering was performed between the binding modes, and a short-time structural optimization and a complete structural optimization were performed based on the binding mode compounds representing each cluster.

In the following example, FLEXOS FlexSIS was used in the docking simulation.
In addition, the structure optimization module of SYBYL (registered trademark) software manufactured by TRIPOS is used for the structure optimization, and the process is stopped after 30 steps of the structure optimization, and the parameters after optimization are calculated. did.

  Further, as a method for evaluating the degree of filling, the hydrophobic interaction energy Ehp shown in the above formula (1) was used as a parameter relating to the binding mode.

Example 1
This example relates to a protein having a PDB ID of 1 adb.

  In this example, when calculating Ehp, σ1 = 4.0 and σ2 = 6.0 in the above equations (1) to (8). Further, εhp = 0.8.

  This method was performed on a protein having a PDB ID of 1 adb. In this example, 50 binding modes were generated, and then the binding modes were classified into 10 clusters shown in Table 1. Table 1 is a table showing data obtained when this method is performed on a protein having a PDB ID of 1 adb.

  As shown in Table 1, clustering resulted in 10 binding modes including a binding mode corresponding to the binding mode of the X-ray structure and a binding mode of the other 9 clusters. In the left column of Table 1, the ID of the binding mode representing each cluster is shown.

When structural optimization was performed for these binding modes for a short time, the IDs of the binding modes were 1adb_Xray, 1adb_050, 1adb_030, 1adb_046, 1adb_010, and 1adb_044. The absolute value of has increased.
Also, an RMSD (root mean-square distance) with a known bonding mode was calculated by an internal module of software used for structure optimization. As a result, it was confirmed that the RMSD is zero, that is, the bond mode corresponding to the bond mode of the X-ray structure is included in the bond mode having an increased absolute value of the filling degree.

(Example 2)
This example relates to a protein whose PDB ID is 1add.

  In this example, when calculating Ehp, σ1 = 4.0 and σ2 = 6.0 in the above equations (1) to (8). Further, εhp = 0.8.

  This method was performed for a protein with PDB ID of 1add. In this example, 50 binding modes were generated, and then the binding modes were classified into 15 clusters shown in Table 2. Table 2 is a table showing data obtained when this method is performed on a protein whose PDB ID is 1add.

  As shown in Table 2, clustering resulted in 15 binding modes, including a binding mode corresponding to the binding mode of the X-ray structure and the binding mode of the other 14 clusters. In the left column of Table 2, the ID of the binding mode representing each cluster is shown.

When structural optimization was performed for these coupling modes for a short time, the absolute value of the degree of filling before and after optimization increased for the six coupling modes whose ID of the coupling mode includes 1add_Xray.
Also, the RMSD with a known binding mode was calculated. As a result, it was confirmed that the RMSD is zero, that is, the bond mode corresponding to the bond mode of the X-ray structure is included in the bond mode having an increased absolute value of the filling degree.

  As shown in Tables 1 and 2, the X-ray structure bonding mode, that is, the actual bonding mode, is obtained in the one whose degree of filling is improved after the structure optimization. That is, it is suggested that an appropriate bonding mode is included in the predicted bonding mode if the one having a good filling degree after short-term structural optimization is selected.

It is a functional block diagram which shows the structure of the screening system in embodiment. It is a functional block diagram which shows the structure of the screening system in embodiment. It is a flowchart explaining the screening method in embodiment. It is a flowchart explaining the screening method in embodiment. It is a flowchart explaining the screening method in embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Screening system 101 Structure information reception part 103 Connection style production | generation part 105 Connection style production | generation part 107 Optimization process part 109 Optimization process part control part 111 Parameter calculation part 113 Selection part 115 Output part 117 Storage part 119 Structure information storage part 121 Coupling mode storage unit 123 Parameter storage unit

Claims (8)

Generating a plurality of binding modes of the low molecular weight compound to the protein by docking simulation of the protein and the low molecular weight compound;
Calculating a parameter relating to the binding mode for the plurality of generated binding modes;
Starting a structure optimization process for the plurality of generated coupling modes;
Stopping the structure optimization process before the structure optimization process is completed;
Calculating the parameters for the plurality of coupling modes after stopping the structure optimization process;
For each coupling mode, comparing the parameters before and after the start of the structure optimization process, and selecting a coupling mode in which a predetermined change has occurred in the parameter from the plurality of coupling modes;
A screening method comprising: The screening method according to claim 1,
The screening method, wherein the parameter is a parameter relating to the degree of packing of the low molecular compound with respect to the protein. The screening method according to claim 2, wherein
In the step of selecting the binding mode, the parameter relating to the degree of packing is changed so that the low molecular weight compound is loaded in the protein after the structure optimization process is stopped and before the start of the structure optimization process. A screening method for selecting a binding mode. The screening method according to claim 2 or 3,
The parameter relating to the degree of packing is a hydrophobic interaction energy between the protein and the low molecular compound,
In the step of selecting a binding mode, the hydrophobic interaction energy value after stopping the structure optimization process is less than the hydrophobic interaction energy value before starting the structure optimization process. Screening method to select the format. A binding state generation processing unit that obtains structural information of proteins and low-molecular compounds, and generates and processes binding modes between the proteins and the low-molecular compounds;
A parameter calculation unit for calculating a parameter related to the coupling mode for the plurality of coupling modes;
Based on the parameters calculated by the parameter calculation unit, a selection unit that selects a coupling mode from the plurality of coupling modes;
Including
The combined state generation processing unit
A binding mode generation unit that generates a plurality of binding modes of the low molecular weight compound to the protein by docking simulation of the protein and the low molecular weight compound;
An optimization processing unit that performs structure optimization processing of the coupling mode generated by the coupling mode generation unit;
An optimization processing unit control unit for stopping the structural optimization processing in the optimization processing unit at a predetermined stage;
Including
The selection unit acquires the parameters before and after the start of the structure optimization process for each coupling mode, compares the parameters before and after the start of the structure optimization process, A screening system for selecting a binding mode in which a predetermined change has occurred from the plurality of binding modes. The screening system according to claim 5, wherein
The screening system in which the parameter calculation unit calculates a parameter relating to the degree of packing of the low molecular compound with respect to the protein in the binding mode. The screening system according to claim 6,
A binding mode in which the parameter relating to the degree of filling is changed so that the low molecular weight compound is filled in the protein after the structure optimization process is stopped before the selection unit starts the structure optimization process. Choose screening system. The screening system according to claim 6 or 7,
The parameter relating to the degree of packing is a hydrophobic interaction energy between the protein and the low molecular compound,
The selection unit selects a binding mode in which the value of the hydrophobic interaction energy after the structure optimization process is stopped is smaller than the value of the hydrophobic interaction energy before the structure optimization process is started. Screening system.

Images