Oxygen tension and sister-chromatid exchanges

Oxygen tension and sister-chromatid exchanges

Mutation Research, 121 (1983) 67-69 67 Elsevier MRLett 0409 Oxygen tension and sister-chromatid exchanges Betsy Hirsch and Jaroslav C e r v e n k ...

116KB Sizes 0 Downloads 8 Views

Mutation Research, 121 (1983) 67-69



MRLett 0409

Oxygen tension and sister-chromatid exchanges Betsy Hirsch and Jaroslav C e r v e n k a Division of Cytogenetics and Cell Genetics, Department of Oral Pathology and Genetics, University of Minnesota, Health Sciences Unit A 16-139, Minneapolis, M N 55455 (U.S.A.) (Accepted 23 March 1983)

In a recently published report, Guti6rrez and L6pez S~iez investigated the effects of oxygen tension on SCE production in Allium cepa L. meristem cells (Guti6rrez and L6pez S~iez, 1982). Their very interesting results clearly show an increase in SCE frequencies as a function of oxygen tension. The authors present a theoretical model to describe and predict the effect of O2 tension on SCE. In this model, the assumption is made that baseline SCE production per polynucleotide chain per cycle is constant for all oxygen tensions. This is represented in their theoretical predictions by the inclusion of the constant 8 × to all equations. Further, the authors conclude that the increase in SCE with increasing 02 tension is dependent solely upon the interaction with BrdU-substituted strands. Our own analyses of Guti6rrez and L6pez S~iez' data have shown that this model does not provide an adequate fit of the observed to the predicted values. We propose another model which may provide a better fit to the data which differs mainly in its provision for the effects of 02 tension on both thymidine and BrdU-substituted strands. A Chi-Square test of goodness of fit (Lindgren, 1976) for the Guti4rrez and L6pez S~iez model can be applied using the results given in their Table 2:


1 2 3 4

1st S period

2nd S period

Expected Observed S.D. of









BrdUrd BrdUrd BrdUrd BrdUrd

20 20 20 20

dThd dThd BrdUrd BrdUrd

20 100 20 100




s(O)a 2.62 3.36 3.78 6.00

2.75 3.59 3.88 5.40

0,097 0.122 0.112 0.113

as(O) was calculated by taking +95% confidence interval of the mean of the observed divided by 1.96.

0165-7992/83/$ 03.00 © 1983 Elsevier Science Publishers B.V,


A Chi-Square can be defined by 4

X2 =



-- E ) 2 / S 2 ( 0 )


Using the data presented above X2 -~ 2 6 . 5 8

Regardless of whether the authors generated the expecteds using parameters estimated on an independent sample (df -- 4) or upon the same sample (dr = 1), the x 2 is significant (p<0.05); i.e., there is a significant difference between the expecteds and the observeds. An alternative model to explain the effect of oxygen tension on SCE is described below. It provides for interactions o f 02 tension with both thymidine- and BrdUsubstituted strands: Let x represent the number of baseline SCE per thymidine-substituted polynucleotide chain per cycle. Let X2o = the number of SCE per thymidinesubstituted chain per cycle under 02 tension 20070; xloo under 02 tension of 100°70. Let y be defined as by Guti6rrez and L6pez S~iez, i.e. y = number of baseline SCE per BrdU-substituted chains per cycle; y2o = SCE per BrdU-substituted chain when O2 tension is 20070; y~oo when O2 tension is 100070. The equations given in the authors' Table 2 m a y be rewritten as follows:




1 2 3 4

5x2o 2X2o 3x2o 2X2o

2.75 3.59 3.88 5.40

+ + + +

3y~o 2y2o + 3xloo + YlOO 5y2o 2y2o + xloo + 3yloo

Solution o f these simultaneous equations will necessarily give an exact fit to the observations. The solutions are: X2o = 0.132 y2o = 0.697

Xloo = 0.257 yloo = 1.162

The standard errors for these estimates can be calculated using s(O)'s presented previously. For example, the standard error of X2o can be calculated as the square root of the variance of X2o which equals: 25 (0.097) 2 + 9 (0.112)2/162 = (0.037) 2 Further,



Standard error

x2o Y2o Xloo yloo

0.037 0.039 0.050 0.053

The validity o f this model can be tested only by obtaining further independent observations. However, we can test the significance of the difference between X2o and X~o0 and y2o and y~oo as ( x l o 0 - X2o) = 0.125 ___ 0.062 ( y l o o - y2o) = 0.465 ___ 0.068 For the difference in x, the Chi Square (x 2 = 4.06) is significant at p < 0 . 0 5 , and for the difference in y, the Chi Square is significant at p = 0 (X2 = 4676.1). Thus, there is evidence for a change in both x and y with increasing 02 tension. In conclusion, we have found that the model proposed by Guti6rrez and L6pez S~tez to account for the increase in SCE with increasing 02 tension does not provide an adequate fit to the observed data. We present an alternative model which permits an interaction of 02 tension with the thymidine-substituted strand as well as with the BrdU-substituted strand. Our model predicts a greater effect of 02 tension on BrdU-substituted strands, but a significant effect with thymidine-substituted strands as well. This model has certain theoretical appeal, but its validity awaits further testing on independent observations.

References Guti~rrez, C., and J.F. L6pez-Sfiez (1982) Oxygen dependence of sister-chromatid exchanges, Mutation Res., 103, 295-302. Lindgren, B.W. (1976) Statistical Theory, 3rd edn., Macmillan, New York.