![Sample our Medicine, Dentistry, Nursing & Allied Health journals, sign in here to start your FREE access for 14 days]({"type":"image" , "src" : "/sda/5944/TOC-Subject-Sample-2021-Medicine-Dentistry-Nursing-Allied-Health.jpg"})
Abstract
We provide detailed protocols for applying the QSAR decision-rule models described in Part 2 of this paper. These procedures permit prediction of the intracellular localization of fluorescent probes or of any small molecular xenobiotic whether fluorescent or not. Also included is a set of notes that give practical advice on various possible problems and limitations of the methods, together with a flow chart that provides a graphical algorithmic summary of the QSAR models.
Notes
Acknowledgments
RWH thanks Dr. R. Aitken, School of Life Sciences, College of Medical, Veterinary and Life Sciences, University of Glasgow, for providing facilities. JCS thanks Dr. M. Cañete and Dr. A. Villanueva for valuable collaboration, and Ministerio de Economía y Competitividad (Spain) for the grant CTQ2013-48767-C03-03).
Declaration of interest: The authors report no conflicts of interest. The authors alone are responsible for the content and writing of this paper.
Notes
1. Chemical structures provided by monographs, handbooks, vendors’ catalogues and research papers may be incorrect or misleading (CitationHorobin and Rashid-Doubell 2013, CitationStockert and Abasolo 2011). Also consider FootnoteNote 2.
2. If this is outside your area of expertise, make friends with a chemist.
3. Micro pKa values are those related to particular ionizable groupings within a probe. Check whether pKa values given for a compound actually refer to this type of value. If compounds have several ionizable moieties, a single “blended” experimental pKa value may be given, which is not what is required here. This issue has been discussed by CitationPerrin et al. (1981, p. 17) and CitationAvdeef (2003).
4. When using information sources, remember that micro values are not always provided and that reported values often have been estimated, not measured.
5. Molecules may contain single conjugated regions or multiple regions separated by non-conjugated units, such as methylene groups.
6. Do not count the non-conjugated bonds that link isolated conjugated rings or chains. Do not count bonds that link conjugated rings or chains in which not all atoms fall on a single plane.
7. Worked examples of the procedure are provided by CitationHorobin and Rashid-Doubell (2013).
8. Procedures such as these vary widely, as do the validities of the outcomes. For more information, read a review by CitationMannhold et al. (2009) that critically discusses and compares outcomes and ease of use of a number of computational methods.
9. Software packages often do not allow estimation of Log P values of ionic species. Even when the structure of an ion is used as the basis of the input, the Log P estimates generated may be those of a corresponding unionized species.
10. Do not assume the Hansch and Leo procedure can be learned in an hour or so; it is complicated.
11. The hydrophilic and hydrophobic zones of small molecules with little conformational flexibility often can be assessed using a two-dimensional structure diagram. More flexible or complex molecules call for inspection of three-dimensional structures. In these cases the plausibility of the estimated AI parameter is reduced.
12. An example often arising with fluorescent probes is that acids can result from hydrolysis of esters, a reaction that is catalyzed by cell esterases.
13. Hardcopy and online documentation often show chemical structures of non-ionized species even when, under physiological conditions, the species present include anionic or cationic molecules. It is common for multiple ionic species to require consideration in the protocol presented here.
14. Estimates are required for each ionized species, because each species is distributed intracellularly as if it were a separate compound (CitationFranco and Trapp 2008).
15. The models specified in Part 2 of this paper assume that probe delivery is by an external solution. If this delivery mode is changed, the localization pattern of a probe may also change. For example, in Part 2, DiI is given as an example of a superlipophilic dye that stains the plasma membrane selectively when applied from an external solution. If DiI is administered directly into the cytosol, however, then other cell membranes, including those of the endoplasmic reticulum, are stained (CitationTerasaki et al. 1994).
16. Owing to the vagaries of cell and protocol factors, no definite definition of “prolonged” is possible; however, it may be taken to be longer than a few minutes.
17. Staining protocols can become ritualized and the function of a procedural step may no longer be appreciated by the bench worker. Therefore, treating stained cells with a probe-free solution, or even with a solution containing an agent intended to complex with the probe, e.g., dextrin or serum albumin in the case of lipophilic probes, can be termed a “washing.” This may misleadingly be interpreted as implying merely the removal of probe that is not yet internalized. Some protocols, however, do indicate the removal of internalized probe by use of terms such as “back exchange” or “extraction.”
18. Section D above is not to be read as suggesting that localization is completely controlled by the physicochemical character of the probe. Cell and protocol factors also influence localization patterns as discussed in Part 1 (CitationHorobin and Stockert 2011).
19. Because the QSAR models involved are dichotomous, a molecular species should fall into just one of these categories. When structure parameter values lie near or on boundaries of key regions in the parameter space of a QSAR model, however, ambiguity of prediction will arise.
20. When multiple species or multiple compounds are present, the variant chemical structures could still have the same intracellular target, although this is not always the case.
21. Setting to zero is required for the predictive algorithm given in to be used unambiguously.