Chemistry Reference
In-Depth Information
treatment will focus on classification schemes that link biological mode of action to
coordination chemistry.
There are several candidate metal classification schemes to employ for SAR
and QSAR generation (Duffus 2002). The easiest to eliminate at the onset is
classification based on whether the metal is an unstable or stable nuclide. This
classification is irrelevant because our intent is not prediction of effects arising
from different types of ionizing radiations. It is prediction of adverse effects from
chemical interaction between metal and organism. Classification based on natu-
ral abundances such as
bulk
,
abundant
, or
trace
elements is unhelpful because
we wish to make predictions for toxicological effects at unnatural, as well as
natural, concentrations. However, there are cases in which natural abundance or
natural occurrence information can provide valuable insight, as exemplified by
the studies of Fisher (1986) and Walker et al. (2007), respectively. Another gen-
eral classification of metals is the dichotomous division of metals as either being
heavy or light metals. The general cutoff between these two groupings (circa 4 g
cm
−3
) has been applied loosely to highlight the toxicity of many heavier metals.
Obviously, a dichotomous schema has minimal utility here, especially for creating
QSARs. At a slightly finer scale, Blake (1884) did note more than a century ago a
correlation between atomic number and metal toxicity. Conforming to the Irving-
Williams series, toxicity to mice increased progressively with atomic numbers
from manganese (atomic number 25, density 7.43) to copper (atomic number 29,
density 8.96) (Jones and Vaughn 1978), but this increase also corresponded with
the progressive addition of d-orbital electrons from [Ar]3d
5
4s
2
to [Ar]3d
10
4s
1
. Such
a scheme based on density or atomic number does not incorporate important peri-
odicities influencing metal toxicity. A schema framed around the periodic table
seems more amenable because metal binding to critical biochemicals can easily
be related to the classic periodicities therein. Certainly, trends in the nature and
occupation of the outer valence shell can be discussed starting from this classic
vantage, e.g., qualities of d- versus s- and p-block elements (Barrett 2002; Walker
et al. 2003). However, this approach requires extension to generate related quan-
titative metrics of binding tendencies. For example, zinc ([Ar]3d
10
4s
2
) was less
toxic in the above progression (Jones and Vaughn 1978) than might have been
anticipated based on atomic number, density, or the number of d-orbital elections
alone. With the maturation of coordination chemistry as a predictive science, rel-
evant quantitative metrics have emerged that combine several metal ion properties
into directly useful metrics. Continuing the example, the empirical softness index
(σ
p
) described later conveniently resolves the inconsistency just noted for zinc
toxicity. These schemes framed on classic periodicity-related binding tendencies
are favored here.
The primary purpose of classifying [metal ions] in (a), or hard, and (b) or soft, is to
correlate a large mass of experimental facts. All the criteria used for the classification
are thus purely empirical; they simply express the very different chemical behavior of
various [metal ions].
Ahrland (1968, p. 118)
