The main conclusions resulting from this project can be summarised as follows:

1. The assay can be conducted at different wavelengths from that previously used, but to prevent a loss in discriminatory power more enzyme or a longer reaction period must be used.

2. Anodonta gill GST is active towards a range of substrates, although this range is not as great as that of rat liver GST. It is probable that this activity is due to multiple enzymes, but the low activity limits the potential for their use in a test.

3. The measurement of Anodonta gill GST can be used to detect pollutants at a concentration which is hazardous to fish, but apparently of low toxicity to mussels.

4. An understanding of the complex relationship between GST, GSH and pollutant exposure has been gained.

5. An assay for determining GSH has been developed which has several advantages over those currently in use. The variablility is high, but is comparable with the variability reported for other methods. However the potential of a GSH based assay for exposure is limited by the transient nature of its modulation.

The demonstration of a clear and logical relationship between exposure and response was the most important aim of the laboratory work. In this respect, the discovery of a linear induction of GST with time at low to moderate exposure to chlorathalonil is very important. An ultimate aim would be to derive a mathematical model to enable determination of pollutant exposure from GST activity. Although there is insufficient data to justify this approach on the basis of the present work, an equation can be derived to demonstrate the principal:

2. Given this then,
gradient for 1 mg/l = gradient at 8 mg/l * 1.73³, since 8 = 2³
= 0.3072

c = 2^x
(log c) / (log 2) = x

So, gradient = 0.3072 * 1.73^x
= 0.3072 * 1.73^{(log c /log 2)}

4. Assuming a baseline activity of 28.5, then the activity (GST) at any day and at any concentration is given as:

GST = 28.5 + 0.3072 * 1.73 ^{(log c / log 2)} * day
and so, log c = 1.26 log [(GST - 28.5) / 0.3072 * day]

Where, GST = activity in nmol/g/sec
day = number of days exposure
c = exposure concentration

This equation gives an activity which approaches 28.5 as exposure concentration and time approach zero. It is, however, undefined when at zero or negative concentrations. Exposure concentration as a function of activity is also undefined on day 0.

From a practical point of view, the most useful estimation which can be determined is that of exposure concentration. Future work could therefore concentrate on providing an estimation of GST levels at a single timepoint (around 14 to 28 days after initiation of exposure). If a rigorous and predictive model can be developed, this would provide a paradigm for future in situ work. Pollution levels could then be expressed as, for example, `chlorothalonil equivalents'.

The problem with this technique is the observed depletion in GST activity at higher exposure concentrations, leading to an inverted-U response to pollution. It would not be possible to ascribe a single observation to the high or low dose part of the curve. After 28 days this may be negated as GSH levels are replenished, however Polak (1992) observed a reduction in Anodonta gill GST after a 28 day deployment dowstream of a sewage outfall. It seems possible that if, as this study suggests, this decrease is due to prolonged depletion of GSH, then this may be associated with overt signs of toxicity. (since glutathione is an absolute requisite for many biochemical pathways). Indeed the toxicity of chlorinated phenols towards Daphnia is correlated with their ability to inhibit GST (LeBlanc, G.A., Hilgenberg, B., Cochrane, B.J., 1988). This leads on to the possibility of using a GST test in conjunction with a physiologically based endpoint.

Johnson et al. (1992) found a negative correlation between Mytilus gill GST and scope for growth (SfG), which is essentially that part of metabolism which is not devoted to tissue maintainance. However there was insufficient data to determine the significance of this relationship. However Polak (1992) found no significant difference in condition index (CI, total tissue weight as a proportion of shell volume) in Anodonta deployed up and downstream of a sewage outfall. Of these two methods, CI may provide the best estimation of toxicity since SfG simply measures the diversion of metabolism to detoxification whereas tissue reserves may be more directly related to health. Unfortunately CI is fairly insensitive and significant changes may only be observed after several months (Veldhuizen-Tsoerkan, M.B., Holwerda, D.A., DeBont, A.M.T., Smaal, A.C., Zandee, D.I., 1991).

The investigation of the relationship between glutathione depletion, GST induction and a physiological test should be investigated. Only a relatively crude measure is needed, providing a simple binary outcome. If a physiological effect is noted, then GST measurements can be related to the high exposure component of putative inverted-U model, where GST is inhibited.

An alternative to a dual study is the repeated measure of GST at various timepoints, since it can be seen from study 6 (see section 9) that the profile of the dose-response varies with time. This proposed approach may be the most attractive since, (1) it means that a single technique can be used throughout the study and, (2) it does not depend on a relationship being found between GST activity and systemic toxicity. Stevens et al. (1991) have stated that a biomarker is most useful when it can be regarded as an indicator of effect as well as exposure (Stevens, D.K., Bull, R.J., Nauman, C.H., Blancato, J.N., 1991). It appears from the studies described here that the time dependent change in GST may fulfil this criterion. If this can be verified then the utility of GST as a biomarker shows great potential.

The ancillary aim of this project, that of streamlining and improving the assay, has been partially achieved with the ammendments described in study 4. One of the major problems is the dependence on specialised equipment in the sample preparation stage. This is essentially due to the high speed centrifugation step. Johnson et al. (1992) suggest this needed to remove interference due to NADH (which absorbs at 340 nm). However this seems unlikely since, (1) free cytosolic NADH is much smaller than GST and so unlikely to be removed by centrifugation that leaves GST in suspension and, (2) NADH is a fairly reactive molecule and would be expected to become oxidised during storage.

Therefore it is likely that any interference which is being removed is due to large cytosolic protein aggregates (such as ribosomes and small microsomes), which contain tryptophan - which will also absorb at 340 nm. It is also possible that Anodonta gills contain a protein bound yellow pigment which causes the interference - some observations suggesting this were made in the course of this project work. It seems likely in actual fact that the simple removal of large microsomes and mitochondria by centrifugation at 12,000g would be sufficient. Indeed this is the force used in recent EC sediment tests - although no increase in activity was detected in these tests (Crane et al., 1993). Centrifugation at lower forces may cause problems since Mukhtar et al. (1981) found that the induction of mitochondrial and microsomal GST is less sensitive than that of cytosolic GST. Pooling the three forms of activity would therefore result in a lower overall induction (as a percentage of control) (Mukhtar, H., Baars, A.J., Breimer, D.D., 1981).

There may be some potential in investigating induction effects with different cell fractions since cytosolic, mitochondrial and microsomal GST can be induced differently (Mukhtar et al., 1981). This effect may have been inadvertantly observed earlier when Boryslawskyj et al. found an induction in Sphaerium GST in response to lindane after centrifugation at 3,000g, but the same group later observed no induction after 100,000g (Boryslawskyj, M., Garrood, A.C., Pearson, J.T., 1988, and Johnson et al., 1992). The measurement of cytosolic vs. membrane bound GST activity may provide a mechanism for identifying the type of induction.