Journal of Bioinformatics, Computational and Systems Biology

Computer-Aided Druggability Detector: A Computerized Apparatus for Quickening Drug Discovery Processes

Download PDF

Published Date: Sep 06, 2018

Computer-Aided Druggability Detector: A Computerized Apparatus for Quickening Drug Discovery Processes

Norbert Nwankwo1*, Ignatius Okafor2, Emmanuel Ibezim3, Ngozika Njoku4, Chinazor Nwankwo5, Chioma Chinasa6, Kennedy Oluigbo7 and Tochukwu Okonkwo6

1Department of Clinical Pharmacy, Faculty of Pharmacy, University of Port Harcourt, Nigeria
2Department of Pharmaceutics and Pharmaceutical Technology, Faculty of Pharmacy, University of Jos, Nigeria
3Department of Pharmaceutics and Pharmaceutical Technology, Faculty of Pharmacy, Faculty of Pharmacy, University of   Nigeria, Nsukka, Nigeria
4Nova Psychiatric Services, Quincy, USA
5Faculty of Medicine and Surgery, University of Nigeria, Enugu Campus, Enugu, Nigeria
6Department of Pharmaceutical Chemistry, Faculty of Pharmacy, University of Port Harcourt, Nigeria
7Department of Clinical Pharmacy, Faculty of Pharmacy, Madonna University, Elele, Nigeria

*Corresponding author: Norbert Nwankwo, Department of Clinical Pharmacy, Faculty of Pharmacy, University of Port Harcourt, Nigeria, E-mail: nobatn@gmail.com

Citation: Nwankwo N, Okafor I, Ibezim E, Njoku N, Nwankwo C, et al. (2016) Computer-Aided Druggability Detector: A Computerized Apparatus for Quickening Drug Discovery Processes. J Bioinf Com Sys Bio 1(1): 101.

 

Abstract

 

Druggability is the process of authenticating the level of fitness or complementariness of therapeutic interventions to their ligands or binding sites such as target proteins. It is a measure of affinity (inter-molecular interaction) as well as physiochemical and structural fitness (intra-molecular interaction) existing between them. Embarking on any design or development of a therapeutic agent without validating their suitability and usability has been considered wasteful. Techniques employed in earlier druggability search include rule-based statistical and machine methods such as Neural Networks (NN) and Support Vector Machine (SVM). Results from these techniques are however predicted rather than calculated. Consequently, Computer-Aided Druggability Detector, a useful tool for designing, developing and validating the suitability and usefulness of therapeutic agents has become needful. Therefore, computer-assisted, digital signal processing techniques such as Information Spectrum Method (ISM), which quantify bio-functionalities, are required.

Using this program, inter-molecular and intra-molecular interactions (druggability) existing between two sets of protein- and peptide-based therapeutic agents and their protein targets are numerically presented. The agents include two antiretroviral agents, Enfuvirtide, Sifuvirtide and their protein target, N-Heptad Repeat (NHR) as well as starter materials used in the design and development of malaria vaccines (P18 and P32). Enfuvirtide provided druggability of 72.33%. This drug is already in use for the management of HIV/AIDS. Sifuvirtide offered 85.42%. Their protein target, NHR exhibited 87.07%. The malaria vaccine peptide candidates, P18 and P32 provided 93.4 and 95.15% respectively. These results are consistent with the clinical outcomes. This tool, which with the assistance of a computer algorithm, identifies and computes druggability is therefore presented here as a computer-aided druggability detector. It can be employed to speed up the drug discovery processes.

Keywords: Digital Signal Processing; Druggable Protein; Druggability; Informational Spectrum Method; Therapeutic Agent

Top ↑

Introduction

 

It has been reported that less than one percent of drug discovery projects get into the market, and that failure at the late stage drug discovery processes, the clinical trials is painful because large resources and time must have been consumed [1]. However, one way of increasing success rate in the process is by evaluating the suitability or fitness, hence usability of any drug molecules or the ligands or target sites for therapeutic intervention.

Druggability has been defined as the tendency for orally bio-available small molecules to bind (inter-molecular interaction) to a given target in an attempt to render therapeutic value [2]. It is a process that employs Pharmaco-kinetics and Pharmaco-dynamics machineries (Structural and Physiochemical, hence intra-molecular interaction) in producing therapeutic effects [1]. It is therefore a measure of the therapeutic significance provided by therapeutic agents or their targets that qualify them for use in therapeutic interventions as a result of affinity (inter-molecular interaction) and complementariness provided by the structural and physicochemical properties that gave rise to the pharmacokinetic and pharmacodynamics characteristics (intra-molecular interaction).

Because preliminary clinical approaches employed in druggability search involve investigations that require the use of therapeutic agents, equipment, animal tissues or target organs, reagents and time [3], the process has gone through series of reformations. Based on the understanding that good knowledge of specific physiochemical and structural characteristics are needed in verifying druggability, its search has been assisted by quick advances in genomics, proteomics, protein structure elucidation, and understanding of the molecular mechanisms of diseases, leading to some successful stories of drugs and drug target discoveries [4]. Part of the rationalization processes involved the use of procedures such as the rule-based (e.g. Rule-of-Five) and machine learning procedures (including Support Vector Machine, SVM and Neural Networks, NN) [4]. The Machine-learning procedures rely on mere forecast and as a result may not completely produce expected results. 

There is therefore the need to engage techniques that not only rationally assess and quantify druggability from sequences but provide and engage appropriate properties of the drug, and adequate basis for determining their suitability. Computerized processes appear appropriate. They have become major tools in the designing and assessment of protein/peptide-based drugs, and have currently become the basic procedures for assessing drug activities. One such computerized procedure is the Digital Signal Processing (DSP) techniques, which include Informational Spectrum Method [3,5-21] and Resonant Recognition Model [22].

Additionally, there is astronomical growth in the protein sequences from which embedded bio-functionalities could be extracted. The understanding that bio-active substances are expressed in the genes/protein residues [23] or target proteins may have accounted for the uncovering of vast protein-coding genes in human proteome and other organisms. There are 20,000-25,000 protein-coding genes in human proteome only. Subsequently, over 30,000 unique protein sequences have been unearthed from these genes [24]. In addition, mathematical investigations of the ailments and associated proteins have advocated that the entirety of the anticipated protein targets in the human genome ranges from 600 to 1500 [4].  These numbers will continue to grow daily. Their assessments therefore call for the deployment of computer-aided procedures.

Initially, it appeared impracticable that computerized processes could help elucidate the bio-functionalities (pharmacological activities, physiochemical and structural features, etc) of bio-active substances including drugs, vaccines, toxicants, etc using their protein sequences (where they are proteins/peptides), or the sequences of their target protein, or the genes/proteins encoding them. However, researchers have demonstrated that bio-functionalities are now decipherable from their protein sequence information using computer-aided procedures [12,18,25]. For example, sequence information of the HIV-1 protease enzyme, the target protein for a non-protein antiretroviral agent, Amprenavir, has been recognized to proffer information on its resistance [10]. Other bio-functionalities unveiled from sequences of the protein include molecular mechanisms to neuro-cardiovascular disease [13], HIV progression to AIDS [20], etc. Similarly, bio-functionalities of other agents such as toxicants (Lead, Nickel, Cadmium, etc), could be extracted from the sequence information of genes or proteins they express [18,23].

Assessing Druggability entails aggregating both inter-molecular and intra-molecular interactions that exist between the therapeutic agents and their target sites [1,2]. Inter-molecular interaction is governed by physiochemical properties represented as bio-recognition and bio-attachments. It engages Long Range Interaction (LRI), which describes interaction space within the range of 5-1000 Armstrong [7,8,10-17]. Inter-molecular interaction engages a molecular descriptor or Amino Acid Scale (AAS) [19] called Electron Ion Interaction Potential (EIIP) [5-22]. Inter-molecular interaction is the first step in every interaction. After the interacting bio-molecules have recognized themselves, attached and bound, intra-molecular interactions take over. On the other hand, intra-molecular interactions engage Short Range Interaction (SRI), which expresses interaction less than 5 Armstrong [5-22]. Intra-molecular interactions may be in a form of physiochemical (e.g. Hydrophobicity using AAS such as WILM950103 [26] or structural (e.g. Alpha-Helix AASs such as BURA740101 [27].

We have earlier recognized that all parameters (AASs) associated with the bio-functionalities involved must be engaged and aggregated in order to obtain the entirety of the biological characteristics [14]. AASs have already been described as a set of 20 numerical values expressing the degree of participation of the 20 essential amino acids that make up protein in each interaction [20]. Over 565 AASs have been determined and deposited in various databases such as [28] and various publications.

By inter-changing the protein sequences with the AAS values using all the AASs involved in the interaction from about 565 that are readily available, and further analyzing them using Discrete Fourier Transform (DFT)-based ISM, the entirety of the bio-functionalities (in this case, druggability) of any therapeutic agent or its protein target could be disemboweled. This is because it has been identified that at one point mutation (representing a single interaction), more than one AAS may be involved [29] and as such, aggregation of the contributions by all AASs engaged from all point mutations is necessary for the entire bio-functionalities [14].

Five hundred and sixty five (565) AASs have been recognized. It is necessary to see each AAS as a representative of a specific mode of interaction. In essence, they need not be compared with others as studies had preliminarily taken EIIP as the best AAS [30]. Out of the 565 AASs, inter-molecular interaction engages one out of the two (2) similar set of EIIP according to report of two different authors, Veljkovic [5-21] and Cosic [22]. The intra-molecular interactions engage the other 563 AASs to unveil the physiochemical and structural properties embedded in the proteins. The vastness of the AASs (565) therefore explains the huge numbers of interactions in our systems, ranging from the simplest interactions that may be found in food digestion to complex drug metabolism carried out by the liver [16]. It is therefore necessary to see each AAS as specific for the interaction it denotes. Engaging all the AASs and sequence information involved in the assessment of the totality of bio-functionalities, using ISM technique is one step that has helped transformed ISM procedure to a tool for computerized bio-investigations.

What is also essential is the procedure engaged in determining the required parameters that are involved in the interactions. Some researchers have employed in-depth study of the mechanisms of action of these therapeutic agents, basically at the atomic level [14-16]. Others have utilized computational procedures. For example, Least Angle Regression with Least Angle Shrinkage Selection Operator (LARS-LASSO) was employed in obtaining the 22 parameters involved in the resistance offered by HIV-1 Protease Enzyme on Amprenavir actions of these therapeutic agents is utilized [29].

We have preliminarily engaged ISM-based computer-aided technique, sequence information and all the parameters involved to uncover and distinguish the therapeutic activities of Enfuvirtide, Sifuvirtide, and their target protein (N-Heptad Repeat) [15,16,21]. Equally, we have studied their suitability for use as therapeutic agents (Druggability) in the management of HIV/AIDS. Druggability that exists between these therapeutic agents and their target proteins was established then. The findings revealed that Sifuvirtide (85.4%) demonstrated higher druggability than Enfuvirtide (72.33%). These results were in accord with the clinically derived outcomes. This is because Sifuvirtide exhibited greater potency than Enfuvirtide using biophysical and CD spectrometeric procedures [15,16,21,31,32].  The target protein, NHR offered 87.07% readiness for interaction (druggability) with the therapeutic agents. Apart from assisting in demonstrating that druggability is decodable from protein sequences, this study also presents a new tool called Computer-Aided Druggability Detector.

What has been worrisome is not only certifying the suitability or fitness, hence usability of any drug for therapeutic intervention (druggability) but from their sequence information [4]. This has led to question such as “can druggable proteins or therapeutic agents be deciphered from their sequences”? [4]. Here, we present a tool, which assesses druggability and achieves from the sequences. This apparatus is envisaged to help quickening drug discovery procedures.

This procedure engaged has fetched four other biomedical/bioengineering devices including Computer-Aided Drug Resistance Calculator (Patent Application Publication No: US20150370964) [14], Computer-Aided Vaccine Potency Assessor (InnoCentive Award ID: 9933477), Computer-Aided Drug Pharmaco-Investigator [3,16], Computer-Aided Drug Potency Decoder [16,17]. By means of the Computer-Aided Drug Resistance Calculator, the reliability of this procedure has been validated using 10 control sequences [14]. The results of the control sequences demonstrated very wide range of resistance (42%-100%), unlike the 295 sequences of the HIV Protease enzyme, which maintained narrower range.

Computer-Aided Druggability Detector is adaptable to all therapeutic agents. To demonstrate this, the result of the investigations on two Malaria peptide vaccine candidates, P18 and P32 is used. The findings made in the studies correlated with the clinically derived outcomes [15,16,21]. The detector should be of assistance in the design and development of malaria vaccine since there are over 4611 peptides identified in a species of Plasmodium falciparium [15,16]. Clinical approaches used in assessing the druggability of these plasmodial peptides can be tedious and challenging.

The Procedure

 

Materials

Informational Spectrum Method (ISM), fourteen (14) amino acids parameters, and the sequences of two anti-retroviral peptide agents (Enfuvirtide and Sifuvirtide) as well as their target protein, NHR (table 1) were engaged. Additionally, two Malaria vaccine peptide candidates, P18 and P32 (table 2) were also engaged. The parameters include EIIP, BURA740101, PONP800104, ARGP820101, PRAM900102, ENGD860101, FASG890101, JURD980101 and WOLR790101. Others are CHAM830107, CHAM830108, FAUJ880111, FAUJ880112 and KLEP840101. The sequences were retrieved from Uniprot [33].

Table 1: Showing the sequences of the druggable proteins/peptides, Enfuvirtide and Sifuvirtide, and their target sites N- Heptad Repeat (courtesy 15,16).

Table 2: Showing the sequences of the therapeutic agents, P18 and P32 (Malaria Vaccine starter materials) (courtesy 15, 16).

Top ↑

Methods

Informational Spectrum Method (ISM) has been detailed [5-11,14-16,19-22]. It is briefly explained in this section. It involves three steps.

Step 1:  The Making of Signals (Numerical Sequences) out of the Alphabetic Codes of the Protein Sequences using the parameters (for example: WOLR790101)

This entails inter-changing the alphabetic codes, representing the protein residues into signals (numerical sequences). As shown in figure 1, the alphabetic codes representing the protein residues of Enfuvirtide are exchanged with their corresponding numerical values of the parameter, WOLR790101 (table 3). This results in a numerical sequence (Signal).

Table 3: Showing the 20 values, which represent the level of participation of the 20 essential amino acids that make up protein using a Hydrophobicity parameter known as WOLR790101 (inventor, Wolfenden et al, 1979).

Figure 1: showing the stepwise transformation of the Alphabetic codes of the protein sequences of the Enfuvirtide (table 2) into Signal or Numerical Sequences using a parameter known as WOLR790101.

 

Step 2:  Application of the Discrete Fourier Transform (DFT)

Having transformed the therapeutic agents and as an example, the anti-retroviral agent, Enfuvirtide protein residues into a signal as a plot of the numerical sequence (figure 1), a Digital Signal Processing procedure was applied to it in same manner they are applied to signals obtained from sound for speech detection or in radar technology, etc. This is in order to uncover the druggability of the Enfuvirtide and Sifuvirtide, and their protein target, the NHR.

Sequel to the application of the Discrete Fourier Transform (DFT), the sequence length of all the proteins are brought to same length (window size) by adding zero to those of proteins that are shorter than the longest sequence. This is called zero-padding.

DFT is expressed as:

  Equation 1

  • Where ∑ is a summation sign and e is for the exponential function.
  • j is the imaginary component of a complex number, the square-root of -1,  i.e. (-1)1/2 or √-1.
  • N is the window size based on the maximum amino acid length and in analyzing of the anti-retroviral agents and their target protein, N is 48 (being the length of the NHR).
  • n refers to the discrete time index running from 0 to N-1 for example, 0, 1, 2, 3, ..., N-1. In the case of the anti-retroviral agents and their target protein, where NHR has the highest length and others need to be zero-padded, it will run from 0 to 47 (N-1).
  • k is the discrete frequency index that runs from 0 to N/2. This is because, DFT results in mirror image and the first half of the discrete frequency index is by principle utilized.
  • x(k)(k) expresses the m member of the signal (numerical sequence).

Top ↑

DFT analysis results in complex components. This consists of real, imaginary and absolute components. The absolute components, which are the square-root of the sum of the squared values of the real and imaginary parts, are usually engaged, though imaginary components have shown to be useful and have been employed in some studies [34].

Absolute values are expressed as:

S(a)(n)=X(n)X*(n)=|X(n)2|

      Equation 2

For a given protein such as Enfuvirtide and Sifuvirtide and their target protein (NHR):

  • S(a)  represents the Absolute Spectrum.
  • X(n) is the DFT coefficient.
  • X*(n) is the conjugate, (the * function in equation 2 symbolizes the presence of a complex component).
  • n=1 to N/2.

The outcome of the DFT processing of the numerical sequences (signal) is a plot called Informational Spectrum (IS). Represented at the x-axis is the frequency and at the y-axis, the amplitude/magnitude of the interaction, which in this case is the druggability of the therapeutic agents or the target sites. For easy assessment of contribution or magnitude of the interaction by individual sequence and parameter, the amplitude, which is presented on the y-axis can be expressed in percentage [22].

Step 3: Common Informational Spectrum (CIS)

This implies the point-wise multiplication of the DFT derived magnitude of interaction and in this case, the therapeutic relevance arising from affinity and fitness (druggability) provided by the anti-retroviral agents, Enfuvirtide and Sifuvirtide on their target protein (NHR). Point-wise multiplication helps demonstrate the point of common contribution as the position of the highest amplitude or magnitude of interaction, from which the level of input by each therapeutic agent and their target protein will be assessed.

Common Informational Spectrum (CIS) is expressed as:

 

Equation 3
Where:

  • ∏ is a Point-Wise Multiplication sign.
  • S(a) is the Absolute of the Common Informational Spectrum.
  • m = 1,2, 3,  ..., M
  • M stands for the number of proteins or peptides involved. In this analysis, they are three (3), namely Enfuvirtide and Sifuvirtide and their target protein (NHR).

Point-wise multiplication of the absolute values of each protein engaged in the analysis demonstrates a maximum peak at the point of common interaction for proteins with common functionalities. This point is called Consensus Frequency (CF) [5-11,14-16,19-22]. Having established the CF, the therapeutic relevance of a peptide-based agent and its target protein can now be presented numerically as contributions from inter- molecular interaction using EIIP parameter, and intra-molecular interaction by means of other parameters engaged.

Results

 

Antiretroviral Agents (Enfuvirtide and Sifuvirtide) and their Protein Target

The results shown here and utilized in demonstrating the functioning of the Computer-Aided Druggability Detector have preliminarily been presented [15,16,21]. Our findings (table 4) revealed that Enfuvirtide provided 72.33% as total interaction from the nine parameters engaged, consisting of 81.5% (inter-molecular interaction) and 72.2% (intra-molecular interaction). Sifuvirtide offered greater druggability than the Enfuvirtide. It provided a total of 85.42% comprising of 100% (inter-molecular interaction) and 83.6% (intra-molecular interaction. This appears to signify that it has higher Druggability than Enfuvirtide (72.33%). This result is in conformity with the clinical outcome. NHR’s readiness was determined as 87.07% representing total Intra-molecular, made up of 77.8% (inter-molecular interaction) and 88.2% (intra-molecular interaction). 

Table 4: Showing druggability in terms of the totality of the inter- and intra-molecular interaction that exist between the druggable proteins/peptides, Enfuvirtide and Sifuvirtide, and their target sites N-Heptad Repeat as derived using Computer-Aided Druggability Detector-Based technique (courtesy 15,16).

Top ↑

Malaria Vaccine Starter Materials (P18 and P32)

As revealed in table 5, druggability or therapeutic relevance of the Malaria vaccine peptide candidates P18 and P32 computationally derived using the Computer-Aided Druggability Detector is significant. Peptide P18 demonstrated 93.4% druggability comprising of 100% inter-molecular interaction as well as the 92.1% intra-molecular interaction. P32 revealed 95.2% druggability that comes from 100% inter-molecular interaction and 94.2% intra-molecular interaction.

Table 5: Showing druggability in terms of the totality of the-inter and intra molecular interaction that exist between the druggable proteins/peptides, P18 and P32 (Malaria Vaccine starter materials) as derived using Computer-Aided Druggability Detector-Based technique (courtesy 15,16).

 

Figures 2-4 are the plots of the point-wise multiplication (Common Informational Spectrum or CIS) of the anti-retroviral agents Enfuvirtide and Sifuvirtide, derived using four AASs. They are EIIP, BURA740101, PONP800104 and ARGP820101. While PONP800104 and ARGP820101 identified position 13 as their CF, EIIP and BURA740101 indicated positions 9 and 18, respectively as their CFs. The plots for parameters PRAM900102, ENGD860101, FASG890101, JURD980101 and WOLR790101 are not shown. However, as shown in table 4, they share common CF at a frequency of 0.283 (position 13).

Figure 2: Showing the Point-Wise Multiplication (Common Informational Spectrum or CIS) of the Enfuvirtide and Sifuvirtide using EIIP  (A) and BURA740101 (B) signifying CF at positions 9 and 18, respectively (courtesy 15,16).

Figure 3: Showing the Point-Wise Multiplication CIS  of the Enfuvirtide and Sifuvirtide using PONP800104  (C) and ARGP820101 (D) identifying position 13 as the CF for both parameters (courtesy 15,16).

The common informational spectrum of the Malaria vaccine peptide candidates, P18 and P32 are also shown in figure 4. The two figures demonstrate the plots for parameter EIIP (figure 4E), which identified position 12 as the CF, and the informational spectrum of the peptide P18 (figure 4F), defining position 12 as point of maximum binding at the CF.

Figure 4: Showing the Point-Wise Multiplication (Common Informational Spectrum) plot of the Malaria Vaccine peptide candidates P18 and P32 using EIIP (E) identifying position 12 as the Consensus Frequency, and the Informational Spectrum of the Peptide P18 defining position 12 as point of maximum binding as also determined by same parameter and presenting  1 or 100% as contribtion by P18 using EIIP (courtesy 11, 12).

 

While parameters CHAM830107 and KLEP840101 share same CF (frequency of 0.094 or position 3), CHAM830108, FAUJ880111 and FAUJ880112 maintained different CFs at frequencies 0.438 (position 14), 0.032 (position 1) and 0.156 (position 5), respectively (table 5).

Top ↑

Discussion

 

Druggability, the entirety of inter-molecular (binding) and intra-molecular interactions (complementariness) as assessed and aggregated using all the parameters and sequence information demonstrated suitability and usability for the five therapeutic agents investigated. The two anti-retroviral agents, Enfuvirtide (72.33%) and Sifuvirtide (85.42%) as well as their target protein, NHR (87.07%) demonstrated high druggability that recommends their use in designing anti-HIV/AIDS drugs though Enfuvirtide is already in use. The two Malaria vaccine peptide candidates, P18 (93.40%) and P32 (95.15%) also exhibited reasonable degree of druggability that would encourage their engagement in the designing of the Malaria vaccines.

Preliminary biophysical analyses of the Sifuvirtide and Enfuvirtide have shown that Sifuvirtide has six fold HIV entry inhibitory activity greater than Enfuvirtide. This outcome was further validated using CD Spectrometry. CD spectrometry identified 93%, 85%, and 0% as alpha-helical content for Sifuvirtide, C34 and Enfuvirtide, respectively. Alpha-helical content, ability to form SHB and also to block other peptides remain the determining factors in the HIV entry inhibitory activity [15,16,21,31,32]. Another study has shown that by exchanging the exposed residues in the six helix bundle so as to build salt bridge, the CHR analogues are engineered to enhance the antiretroviral activities, through improving their solubility, hence druggability. Sifuvirtide is one of the products of the bio-engineering [31]. In this study, it displayed greater druggability than the Enfuvirtide.

Table 6 illustrates description of the parameters, categories of interactions (inter-molecular, structural and physiochemical) and their input, contributions from the therapeutic agents using various parameters, and the totality of the interaction (druggability). While the inter-molecular interactions employ only one parameter, EIIP, the intra-molecular interaction engages more than one parameter. As a result, aggregation of the contributions by various parameters of the intra-molecular interaction is necessarily achieved. Table 6 demonstrates that the structural contributions to the Druggability of Enfuvirtide remain 46.7%, 74.5% and 64.8%. The contributions came in the form of “Normalized frequency of alpha-helix” (BURA740101), “Surrounding hydrophobicity in alpha-helix” (PONP800104) and “Relative frequency in alpha-helix” (PRAM900102), respectively. Table 6 also indicated that the major physicochemical contributions came from Hydrophobicity indices.

Table 6: Showing the names of the inventor of the parameters, year of publication, names of the parameters, descriptions of the parameters, categories of interactions, and contributions from Enfuvirtide, Sifuvirtide, and NHR by various parameters. About 544 sets of 20 numerical values of these parameters are available at www.genome.jp/aaindex. [19].

Top ↑

It has been reported that the HIV entry process is incomplete, and therefore impossible without structural (helical) re-arrangement of N-terminal Repeat (NHR) and C-terminal Repeat (CHR) of HIV gp41 and subsequent pulling of the CD4, perforation and transfer of viral genetic content into the host CD4 [35]. Six-Helix-Bundle (Coiled-Coil) formation must be attained through the conformational (helical) re-arrangement [31]. Disruption of this process disallows contact between the HIV fusion peptide and the host CD4, resulting in the cessation of both entry procedure and infection. Inhibitors of this process are therefore good agents in the management of HIV/AIDS. Currently, there are two antiretroviral agents as well as several other candidates designed out of the CHR and NHR domains of the HIV gp41. Sifuvirtide is still at the clinical trial stage [31].

Enfuvirtide and Sifuvirtide are structural analogues of the CHR. Like the sulfonamides (antibacterial agents), which are structural analogues of Para-Amino Benzoic Acid (PABA) and are incorporated into the erroneous microbial protein synthesis that result in the microbial death, Enfuvirtide and Sifuvirtide, the CHR analogues take over the functions of the CHR in the formation of the false Six Helix Bundle (SHB)  with the NHR (figure 5). This results in the false six helix bundle structure and subsequent cessation of viral entry process. This knowledge, and the reason why the Enfuvirtide and Sifuvirtide, the CHR analogues take over the function of the CHR are fundamental in understanding the druggability of the three agents, Enfuvirtide, Sifuvirtide and their target protein NHR.

Figure 5: Showing vertical and aerial views of the cones representing the helices of the NHR and CHR in the formation of a true Six Helix Bundle (Coiled-Coil), and NHR and Enfuvirtide (T20) in the formation of a false Six Helix Bundle (Coiled-Coil), demonstrating the Competitive Antagonistic Mechanism of the Inhibitory Activity of the Enfuvirtide, Sifuvirtide on the N-Heptad Repeat.

Figure 5 is the vertical and aerial views of the cones representing the helices of the NHR and CHR involved in the formation of a true and false SHBs (Coiled-Coil). In the formation of the true SHB, NHR and CHR are involved. On the other hand, NHR and the CHR analogues (Enfuvirtide, also called T20 or Sifuvirtide) are engaged in the formation of false SHB. As analogues, they act as competitive antagonists to the CHR by taking over the function of the CHR in the formation of SHB with the NHR. This is simply because CHR and their analogues, Enfuvirtide, Sifuvirtide share structural similarity. Each is incorporated in the formation of the false SHB. Except for factors such as affinity (inter-molecular interaction) and complementariness (intra-molecular interaction), it is impracticable for the CHR analogues to displace CHR and take over its function in the formation of SHB.

In this study, the inter-molecular and intra-molecular interactions (druggability) for the two CHR analogues (Enfuvirtide and Sifuvirtide) and the contribution of the NHR are assessed, aggregated and engaged in the demonstration of the functioning of the Computer-Aided Druggability Detector. This tool is anticipated to speed up the drug discovery process. All the agents investigated exhibited high levels of Druggability that suggest their suitability for engagement in the designing of anti-retroviral.

Top ↑

Conclusion

 

By definition, druggability is the aggregation of both inter-molecular and intra-molecular interactions that exist between the therapeutic agents and their target sites [1,2]. Using DFT-based procedure (ISM) and the two sets of therapeutic agents, percentage contribution of inter-molecular by each sequence is calculated using EIIP parameter. Percentage contributions for intra-molecular interactions are also obtained using other parameters that are involved in the structural and physiochemical interactions. These are aggregated to obtain their therapeutic relevance (druggability). The results are found to be in accord with the clinical outcomes. Having calculated and demonstrated this bio-functionality (druggability) in numerical terms with the assistance of computer-based program, we present a tool to be called computer-aided druggability detector.

The potential usefulness of a prediction-free, rational, biomedical and bioengineering tool, which could quicken the drug discovery process, is presented. This tool, which is capable of detecting the affinity and complementariness that exist between therapeutic agents and their protein targets, essentially from their sequence information, is envisaged that it will help quicken the drug discovery processes.

Apart from the fact that druggability search has been saddled with wastefulness in resources and time, the major challenge became how to assess it from sequence information. Using the results of the preliminary investigations [15,16], we demonstrated that assessment of druggability from sequence information is feasible. While we recommend the establishment of cut-off for all drugs and engagement of the device in assessing of the druggability of all other peptide, it becomes intriguing that the resource and time consuming drug discovery processes could be salvaged using Computer-Aided Druggability Detector that require only sequence information only. 

Top ↑

References

 

  1. Edfeldt FN, Folmer RH, Breeze AL.  Fragment screening to predict druggability (ligandability) and lead discovery success. Drug Discov Today. 2011;16(7-8):284-7. doi: 10.1016/j.drudis.2011.02.002.
  2. Yuan Y, Pei J, Lai L. Binding site detection and druggability prediction of protein targets for structure-based drug design. Curr Pharm Des. 2013;19(12):2326-33.
  3. Nwankwo N. Computer-Aided Pharmaco-Investigator: A Biomedical/Bioengineering Device for Pharmacological Investigations. Austin Virol and Retrovirology. 2015;2(2):1015.
  4. Zheng CJ, Han LY, Yap CW, Ji ZL, Cao ZW, Chen YZ. Therapeutic targets: progress of their exploration and investigation of their characteristics. Pharmacol Rev. 2006;58(2):259-79.
  5. Veljkovic V, Niman HL, Glisic S, Veljkovic N, Perovic V, Muller CP. Identification of hemagglutinin structural domain and polymorphisms which may modulate swine H1N1 interactions with human receptor.  BMC Struct Biol. 2009;9(62):1-11. doi: 10.1186/1472-6807-9-62.
  6. Veljkovic V, Veljkovic N, Muller CP, Müller S, Glisic S, Perovic V, et al. Characterization of conserved properties of hemagglutinin of H5N1 and human influenza viruses: possible consequences for therapy and infection control. BMC Struct Biol. 2009;9:21. doi: 10.1186/1472-6807-9-21.
  7. Glisic S, Veljkovic N, Jovanovic Cupic S, Vasiljevic N, Prljic J, et al. Assessment of hepatitis C virus protein sequences with regard to interferon/ribavirin combination therapy response in patients with HCV genotype 1b. Protein J. 2012;31(2):129-36. doi: 10.1007/s10930-011-9381-6.
  8. Veljkovic V, Cosic I, Dimitrijevic B, Lalovic D. Is it possible to analyze DNA and protein sequences by the methods of digital signal processing? IEEE Trans Biomed Eng. 1985;32(5):337-41.
  9. Veljkovic V, Loiseau PM, Figadere B, Glisic S, Veljkovic N, Perovic VR, et al. Virtual screen for repurposing approved and experimental drugs for candidate inhibitors of EBOLA virus infection. F1000Res. 2015;4:34. doi: 10.12688/f1000research.6110.1.
  10. Veljkovic V, Glisic S, Veljkovic N, Bojic T, Dietrich U, Perovic VR, et al. Influenza vaccine as prevention for cardiovascular diseases: possible molecular mechanism. Vaccine. 2014;32(48):6569-75. doi: 10.1016/j.vaccine.2014.07.007.
  11. Veljkovic V, Glisic S, Muller CP, Scotch M, Branch DR, Perovic VR, et al. In silico analysis suggests interaction between Ebola virus and the extracellular matrix. Front Microbiol. 2015;6:135. doi: 10.3389/fmicb.2015.00135.
  12. http://www.vin.bg.ac.rs/180/istree/references.php.
  13. Tijana Bojic, Vladimir R. Perovic, Sanja Glisic. In silico Therapeutics for Neurogenic Hypertension and Vasovagal Syncope. Front Neurosci. 2015;9:520. doi: 10.3389/fnins.2015.00520.
  14. Nwankwo NI, inventor; NWANKWO NORBERT IKECHUKWU, assignee. Computer Aided Drug Resistance Calculator Calculating Drug Resistance Using Amprenavir as a Case Study. United States Patent Publication No: 20150370964 A1. 2015.
  15. Nwankwo N. Signal processing-based Bioinformatics methods for characterization and identification of Bio-functionalities of proteins. 2012;Openthesis.
  16. Nwankwo N, Molokwu G, Njoku N. Novel Computerized Approaches to Investigating Pharmacological Activities”. Computational Biology and Bioinformatics. Computational Biology and Bioinformatics. 2015;3(4):52-64.
  17. Nwankwo N. Computer-Aided Drug Potency Decoder: A Biomedical/Bioengineering Device for Determining and Distinguishing Drug Efficacies. J Bioengineer & Biomedical Sci. 2015;6:176. doi:10.4172/2155- 9538.1000176.
  18. Nwankwo N. Can Bio-functionalities be deciphered from Protein Sequence Information using Computational Approaches? 2015;29th Annual Symposium of The Protein Society Abstract Book:407.
  19. Nwankwo N. A Digital Signal Processing-based Bioinformatics Approach to Identifying the Origins of HIV-1 non B subtypes infecting US Army Personnel serving abroad. Curr HIV Res. 2013;11(4):271-80.
  20. Nwankwo N, Seker H. HIV Progression to AIDS: Bioinformatics Approach to Determining the Mechanism of Action". Curr HIV Res. 2013;11(1):30-42.
  21. Nwankwo N, Seker H. Digital Signal Processing Techniques: Calculating the Biological Functionalities. J. Proteomics Bioinform.2011; 4:260-68. doi:10.4172/jpb.1000199.
  22. Cosic I.  Macromolecular bioactivity: is it resonant interaction between macromolecules?--Theory and applications. IEEE Trans Biomed Eng. 1994;41(12):1101-14.
  23. Lin H, Burton D, Li L, Warner DE, Phillips JD, Ward DM, et al. Gain-of-function mutations identify amino acids within transmembrane domains of the yeast vacuolar transporter Zrc1 that determine metal specificity. Biochem J. 2009;422(2):273-83. doi: 10.1042/BJ20090853.
  24. International Human Genome Sequencing Consortium. Finishing the euchromatic sequence of the human genome. Nature. 2004;431(7011):931-45.
  25. Obermeier M, Ehret R, Berg T, Braun P, Korn K, Muller H, et al. Genotypic HIV-coreceptor tropism prediction with geno2pheno [coreceptor]: differences depending on HIV-1- subtype. Journal of the International AIDS Society. 2012;15(Suppl 4):18214.
  26. Wilce MC, Aguilar MI, Hearn MT. Physicochemical Basis of Amino Acid Hydrophobicity Scales: Evaluation Of Four   New Scales Of Amino Acid Hydrophobicity Coefficients Derived from RP-HPLC of Peptides. J Anal Chem. 1995;67(7):1210-19.
  27. Burgess AW, Ponnuswamy PK, Scheraga HA. T Analysis of conformations of amino acid residues and prediction of backbone topography in proteins. J Isr. J. Chem. 1974;12:239-86.
  28. Tomii K, Kanehisa M. Analysis of amino acid indices and mutation matrices for sequence comparison and structure prediction of proteins. Protein Eng. 1996;9(1):27-36.
  29. Hoj L, Kjaer J, Winther O, et al. In silico identification of physicochemical properties at mutating position positions relevant to reducing susceptibility to amprenavir. XVII International HIV Drug Resistance Workshop. 2008;Poster No.113.
  30. Lazovic J.  Selection of amino acid parameters for Fourier transform-based analysis of proteins. Comput Appl Biosci. 1996;12(6):553-62.
  31. He Y, Xiao Y, Song H, Liang Q, Ju D, Chen X, et al. Design and evaluation of sifuvirtide, a novel HIV-1 fusion inhibitor. J Biol Chem. 2008;283(17):11126-34. doi: 10.1074/jbc.M800200200.
  32. Cai L, Shi W, Liu K. Peptides and Peptidomimetics as Tools to Probe Protein-Protein Interactions –Disruption of HIV-1 gp41 Fusion Core and Fusion Inhibitor Design. Biochemistry. 2012. doi: 10.5772/34034.
  33. Jain E, Bairoch A, Duvaud S, Phan I, Redaschi N, Suzek BE , et al. Infrastructure for the life sciences: design and implementation of the UniProt website. BMC Bioinformatics. 2009;10:136. doi: 10.1186/1471-2105-10-136.
  34. Chrysostomou C, Seker H, Aydin N, Haris PI. Complex resonant recognition model in analysing influenza a virus subtype protein sequences. IEEE  Xplore Digital Library. 2010;1-4. doi: 10.1109/ITAB.2010.5687621.
  35. Menendez-Arias L. Molecular basis of Human Immunodeficiency Virus drug resistance: an update. Antiviral Res. 2010;85(1):210-31. doi: 10.1016/j.antiviral.2009.07.006.

Top ↑

Copyright: © 2016 Nwankwo N, et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.