J. Xu et al., PREFERENTIAL FORMATION AND REPAIR OF CHROMIUM-INDUCED DNA-ADDUCTS ANDDNA-PROTEIN CROSS-LINKS IN NUCLEAR MATRIX DNA, Carcinogenesis, 15(7), 1994, pp. 1443-1450
The distributions of chromium-DNA adducts and DNA-protein crosslinks i
nduced by treatment of intact CHO cells with carcinogenic chromium wer
e examined in distinct chromatin subfractions: a chromatin subfraction
released by digestion of isolated nuclei with micrococcal nuclease (1
SF, 14% of total nuclear DNA), bulk chromatin (74% of total DNA) and a
nuclear matrix fraction (12% of total DNA). The identity of the matri
x fraction was confirmed by hybridization of DNA from each subfraction
with a cDNA probe prepared from total mRNA isolated from CHO cells, w
hich showed that the 1SF and nuclear matrix fractions were 2.3- and 3.
8-fold enriched in actively transcribed genes respectively, compared t
o total unfractionated DNA. Immediately following treatment of cells w
ith 150 mu M sodium chromate for 2 h the binding of chromium to each c
hromatin fraction was found to be non-uniform. Compared with total unf
ractionated nuclei, the nuclear matrix fractions were enriched in chro
matin-bound chromium (3.4-fold), whereas the bulk chromatin fraction w
as relatively depleted (0.5-fold). Approximately 13% of nuclear chromi
um was associated with the detergent-soluble lipid component of nuclei
. A similar distribution of chromatin-bound chromium was also apparent
24 h after the chromate treatment. Immediately after the 2 h chromate
treatment, chromium-DNA adducts were detected in all the chromatin su
bfractions. Total nuclear and bulk chromatin DNA contained similar lev
els of this type of damage. The 1SF fraction was depleted approximatel
y 3-fold in this type of damage compared with total nuclear DNA. In co
ntrast, the nuclear matrix was markedly enriched in chromium-DNA adduc
ts (approximately 4-fold compared with total nuclear DNA) at this time
. As previously demonstrated, chromium-DNA adducts in total nuclear DN
A decreased within the first 24 h, but thereafter persisted at a simil
ar level. Chromium-DNA adducts in nuclear matrix DNA also reached maxi
mum levels at the end of the 2 h treatment and decreased to 68% and 39
% of this level by 24 and 48 h after treatment respectively. In contra
st, the adduct levels in the 1SF and bulk chromatin fractions did not
change up to 48 h after treatment. Chromium-induced DNA-protein crossl
inks, which were stable to 8 M urea and 2% SDS, occurred almost exclus
ively in the nuclear matrix fraction. The crosslinks in this fraction
reached a maximum level at the end of the 2 h treatment, but returned
to control levels 24 h later. Because cell division was completely blo
cked and no redistribution of chromatin was detectable, the loss of th
ese lesions probably represents true repair and the residual persisten
t adducts represent repair-resistant lesions. Treatment of cells synch
ronized in G(1) of the cell cycle and in vitro treatment of isolated n
uclei,vith trivalent chromium both yielded preferential adduct formati
on in nuclear matrix DNA, indicating that the non-random distribution
of lesions was due to a structural predisposition of this chromatin su
bfraction and not accumulation of arrested replication forks or transc
ription bubbles in the matrix. These results indicate that chromium-in
duced DNA damage is both formed and repaired preferentially in nuclear
matrix DNA.