A spectrophotometric assay of the reaction of glutathione with JS-K (Fig. 1) was based on the formation of DNP-SG measurable at 340 nm, as in the reaction with 1-chloro-2,4-dinitrobenzene (CDNB) described by Clark et al. and commonly used to monitor GST activity 9. The leaving group in the conjugation of JS-K is 1-[(4-ethoxycarbonyl)piperazin-1-yl]diazen-1-ium-1,2-diolate, which rapidly decomposes into 4-ethoxycarbonylpiperazine and the projected two active molecules of NO 5. The human genome encodes up to 17 genes of cytosolic (soluble) GSTs, but two of them are frequently represented by null alleles, GSTM1*0 and GSTT1*0, such that the affected individuals do not express the corresponding enzymes 10. Five of the seven classes of cytosolic GSTs are primarily responsible for detoxication of drugs and other xenobiotics: alpha (A), mu (M), pi (P), sigma (S), and theta (T). In the present investigation the classes were represented by human GST A1-1, GST A2-2, GST A3-3, GST A4-4, GST M1-1, GST M2-2, GST M4-4, GST M5-5, GST P1-1, GST S1-1, and GST T1-1.
Table 1 shows the activities with JS-K expressed per mg GST protein. For comparison, the specific activities of the enzymes with CDNB as substrate are shown. All enzyme tested except GST T1-1 demonstrated measurable activity with JS-K. Notably, human GST T1-1 also has negligible catalytic activity with CDNB 8. The JS-K activities of the various GSTs span an approximately 400-fold range.
GST M2-2 displayed the highest specific activity of the tested enzymes: 273 µmol min-1 mg-1 (Table 1). Two other mu class members, GST M5-5 and the polymorphic GST M1-1, gave specific activities of 58 and 57.8 µmol min-1 mg-1, respectively. GST M4-4, which generally has low catalytic activities 11, showed the lowest value of all in the class.
Among the alpha class members, GST A1-1 had the highest specific activity: 42.5 µmol min-1 mg-1 followed by GST A2-2 and GST A3-3 with 37.2 and 18.3 µmol min-1 mg-1, respectively. The lowest activity in the alpha class was shown by GST A4-4, which is also known to have modest CDNB activity 12.
The sigma class enzyme GST S1-1 presented a specific activity of 12.9, twice the GST A4-4 value of 6.3 µmol min-1 mg-1, and likewise has low CDNB activity 13. GST S1-1 is distinguished as the hematopoietic prostaglandin D2 synthase (HPGDS).
The pi class member GST P1-1 ranks among the least active GSTs with a specific activity of 1.2 µmol min-1 mg-1, which is significant because this enzyme is present in most tissues 19 (but not in the hepatocytes of the liver14).
Specific activities have previously been reported for commercial enzyme preparations of GST M1-1 (3.7 µmol min-1 mg-1), GST A1-1 (14.9 µmol min-1 mg-1), and GST P1-1 (0.15 µmol min-1 mg-1) at pH 6.5 and 25°C 5, but later redetermined with somewhat altered conditions of pH 7.4 and 37°C to the higher values of 66.3, 126, and 1.36 µmol min-1 mg-1, respectively 15 The latter values are in general agreement with the values in Table 1 considering differences in reaction temperature and other conditions. On the other hand, the kcat/Km of 20 mM-1s-1 for GST A1-1 reported by the same group is two orders of magnitude lower than our value of 2000 mM-1s-1 in Fig. 2. Our value of 10,000 mM-1s-1 for GST M2-2 is even higher. The mu class GST M1-1 was reported earlier as 63 mM-1s-115.
In general, the activities with JS-K were highly correlated with the activities with CDNB (Fig. 3). This relationship can be rationalized by the similar chemical mechanism, involving an aromatic nucleophilic attack by the sulfur of glutathione to form a Meisenheimer σ-complex, which is followed by the release of the common conjugate S-2,4-dinitrophenylglutathione. A notable exception was the low JS-K activity of GST P1-1 in contrast to the high CDNB activity of the same enzyme. Apparently, the second product moiety of JS-K suppresses the release of S-2,4-dinitrophenylglutathione. In the opposite direction, GST S1-1 deviates somewhat from the general correlation, by being somewhat more active with JS-K than with CDNB (Fig. 3).
The antiproliferative property of JS-K was assumed to be based on NO as the effective agent released from the prodrug 5. An alternative possibility involving S-2,4-dinitrophenylation of thiols or other cellular nucleophiles was excluded based on the lack of an equal effect of CDNB, which could produce a similar arylation without release of NO. However, the finding that pretreatment of HL-60 human myeloid leukemia cells with buthionine sulfoxide, an inhibitor of glutathione biosynthesis, did not suppress the antineoplastic effect of JS-K suggests that other mechanisms than the glutathione-activated NO discharge are involved 5. Accumulating evidence point to the role of signaling through mitogen-activated protein kinases, cell factor β-catenin/T, and ubiquitin-proteasome pathways in cancer cell apoptosis 16.
Even though the molecular mechanisms underlying the therapeutic potential if JS-K remain unresolved, GSTs are unmistakably potent activators of NO release from the compound. The specific activity of 273 µmol min-1 mg-1 displayed by GST M2-2 falls in the domain of the highest activities with any of a wide variety of alternative chemical reactions catalyzed by the enzyme 17. The catalytic efficiency kcat/Km of 1.0x107 M-1 s-1 is approaching the efficiencies of the most efficient enzymes 18 . Two additional mu class enzymes, GST M1-1 and GST M5-5, show 20% of the GST M2-2 activity, and the alpha class members GST A1-1 and GST A2-2 are almost as active (Table 1).
The GSTs as an ensemble of enzymes occur abundantly in all tissues 19-21, and it would appear that the GSTs collectively provide the means for efficient metabolism of JS-K. GST M2-2 is not expressed in liver, but the highly active GSTs A1-1 and A2-2 and GST M1-1 (except in subjects genetically homozygously GSTM1 null) are highly abundant hepatic enzymes such that the tissue has high capacity for the biotransformation of JS-K. Particularly high activity can be expected also in testis, small intestine, kidney, adrenal based on the high expression levels of the latter enzymes. By contrast, uterus reportedly expresses only GST P1-1 and GST T1-1, and like erythrocytes, therefore has comparatively low activity with JS-K.
Overall, it would appear that the abundance of GSTs and their predominantly high activity with JS-K would cause the rapid biotransformation of the prodrug in most tissues. Even though many tumors overexpress GST P1-1 22-24 it is not obvious that the elevation is sufficient to compensate for the relatively low activity of this enzyme in comparison with other GSTs and make the neoplastic cells selectively vulnerable to release of NO from JS-K. In some tumors elevated levels of alpha and mu class GST have also been reported, but neither in these instances can the antineoplastic activity of JS-K be unambiguously attributed to GST-mediated release of NO.
We conclude that the antitumoral effect of JS-K must involve mechanisms more than NO release, and that the interplay between them and the unmistakable degradation of JS-K catalyzed by GSTs requires further incisive investigation.
Table 1. Specific activities of human GSTs with JS-K as substrate.
Assays were made spectrophotometrically at 30°C with 50 µM JS-K and 1 mM GSH in 10 mM Tris-HCl pH 7.5, 1.0 mM EDTA and 0.1% bovine serum albumin. All measurements were made in triplicate and corrected for background reaction. S.D. (standard deviation), Nd (not detectable).
GST
|
Specific activity
(µmol · mg-1 · min-1)
|
±S.D.
|
A1-1
|
42.5 |
±3.5
|
A2-2
|
37.2 |
±0.6
|
A3-3
|
18.3 |
±0.5
|
A4-4
|
6.3 |
±0.2
|
M1-1
|
57.8 |
±2.3
|
M2-2
|
273.0 |
±4.7
|
M4-4
|
0.89 |
±0.02
|
M5-5
|
58.0 |
±0.6
|
P1-1
|
1.2 |
±0.05
|
S1-1
|
12.9 |
±0.01
|
T1-1
|
Nd
|
|