58
3 Single Molecule Non-cleavable Multiply Active Antibacterials
flexible although this came at the cost of predicted accumulation in Gram-negative
bacteria through the increased number of rotatable bonds with increased flexibility.
Another interesting computer aided de novo general design approach to multitargeting is described in work covering the LigBuilder V3 program (Yuan et al.
2020a), but with this approach the target structure is still required. However it is a
promising program which can be used to design ligands to target multiple receptors
or targets, multiple binding sites of the one receptor, or a number of conformations
of one receptor. This program was employed to propose a new compound which
should inhibit both HIV protease and HIV transcriptase at sub-micromolar activity
levels. Application to bacterial target proteins of known structure should be feasible,
and if confirmed, the best resulting designed ligands could also be assessed and
further refined through the web application eNTRyway for likely accumulation in
Gram-negative bacteria (Escherichia coli).
Natural products as starting points
As noted earlier, multi-targeting is the new paradigm in drug design and natural
products form a good starting point as they are often promiscuous in their activity
as noted by Chai and colleagues in their good review on the topic (Ho et al. 2018).
Features of promiscuous compounds include commonly found pharmacophores and
reactive functional groups, typically electrophilic moieties, and often such promiscuity is a problem as in the high throughput screening domain (Baell and Holloway
2010; Baell and Nissink 2018). The problem of nuisance sub-structures in drug leads
also extends to some natural products (Baell 2016), but such compounds could still
be useful as possible starting points for the ultimate development of a drug.
But how to design selective multi-targeting compounds which are not indiscriminately active is the key question here. In other words one would need to reduce the
scope of the promiscuity while retaining controlled multi-targeting. There is a major
problem though achieving selectivity for bacteria over host cells with this approach
and ways to design-out unwanted off target interactions would need to be possible
from modification of the core structure or functional groups in the natural product
starting point, as in the case of berberine and analogues.
High throughput screening
Initial leads to mult-targeting antibacterials can also be accessed via high throughput
screening (HTS) methodology based on large structurally-diverse compound
libraries. While this approach may not be as appealing scientifically perhaps as more
rational approaches, high throughput screening has resulted in some interesting new
antibacterial leads representing new chemotypes. A good recent example of this
was that reported by Ivanenkov et al. (2019a). They discovered a new 2-(pyrazol1-yl)-thiazole–based chemotype using semi-automated HTS with a double-reporter
system (pDualrep2) which identifies molecules capable of blocking the bacterial
translational process or of inducing a response to DNA damage (SOS response).
A large and diverse library of 125,000 compounds was screened. Some of these
compounds displayed high activity against TolC deficient Eschericia coli; TolC is an
important component of the AcrAB-TolC multidrug efflux pump in Eschericia coli.
3 Single Molecule Non-cleavable Multiply Active Antibacterials
flexible although this came at the cost of predicted accumulation in Gram-negative
bacteria through the increased number of rotatable bonds with increased flexibility.
Another interesting computer aided de novo general design approach to multitargeting is described in work covering the LigBuilder V3 program (Yuan et al.
2020a), but with this approach the target structure is still required. However it is a
promising program which can be used to design ligands to target multiple receptors
or targets, multiple binding sites of the one receptor, or a number of conformations
of one receptor. This program was employed to propose a new compound which
should inhibit both HIV protease and HIV transcriptase at sub-micromolar activity
levels. Application to bacterial target proteins of known structure should be feasible,
and if confirmed, the best resulting designed ligands could also be assessed and
further refined through the web application eNTRyway for likely accumulation in
Gram-negative bacteria (Escherichia coli).
Natural products as starting points
As noted earlier, multi-targeting is the new paradigm in drug design and natural
products form a good starting point as they are often promiscuous in their activity
as noted by Chai and colleagues in their good review on the topic (Ho et al. 2018).
Features of promiscuous compounds include commonly found pharmacophores and
reactive functional groups, typically electrophilic moieties, and often such promiscuity is a problem as in the high throughput screening domain (Baell and Holloway
2010; Baell and Nissink 2018). The problem of nuisance sub-structures in drug leads
also extends to some natural products (Baell 2016), but such compounds could still
be useful as possible starting points for the ultimate development of a drug.
But how to design selective multi-targeting compounds which are not indiscriminately active is the key question here. In other words one would need to reduce the
scope of the promiscuity while retaining controlled multi-targeting. There is a major
problem though achieving selectivity for bacteria over host cells with this approach
and ways to design-out unwanted off target interactions would need to be possible
from modification of the core structure or functional groups in the natural product
starting point, as in the case of berberine and analogues.
High throughput screening
Initial leads to mult-targeting antibacterials can also be accessed via high throughput
screening (HTS) methodology based on large structurally-diverse compound
libraries. While this approach may not be as appealing scientifically perhaps as more
rational approaches, high throughput screening has resulted in some interesting new
antibacterial leads representing new chemotypes. A good recent example of this
was that reported by Ivanenkov et al. (2019a). They discovered a new 2-(pyrazol1-yl)-thiazole–based chemotype using semi-automated HTS with a double-reporter
system (pDualrep2) which identifies molecules capable of blocking the bacterial
translational process or of inducing a response to DNA damage (SOS response).
A large and diverse library of 125,000 compounds was screened. Some of these
compounds displayed high activity against TolC deficient Eschericia coli; TolC is an
important component of the AcrAB-TolC multidrug efflux pump in Eschericia coli.