J. Parrington et al., THE INTERFERON-STIMULABLE RESPONSE ELEMENTS OF 2 HUMAN GENES DETECT OVERLAPPING SETS OF TRANSCRIPTION FACTORS, European journal of biochemistry, 214(3), 1993, pp. 617-626
We have previously reported three types of DNA-protein complexes, form
ed specifically with the interferon-stimulable response elements (ISRE
) in the 5' flanking DNA of the interferon-inducible 6-16 and 9-27 gen
es, a type-I interferon-inducible early complex involving factor E (IS
GF3), M and G complexes induced more slowly in response to type-I and
type-II interferons, respectively and CI/C2, a constitutive complex(s)
. Similar complexes have been reported by others. The operationally de
fined band-shift complexes M, G and CI/C2 are shown here to be heterog
eneous and to differ in their factor content, depending on the ISRE pr
obe. With a 9-27 ISRE probe the M, G and C1/C2 complexes all contain t
he gamma subunit of ISGF3, which is present constitutively but is indu
ced in response to IFN-alpha (to yield M) or IFN-gamma (to yield G). I
n contrast, a 6-16 ISRE probe forms band-shift complexes with IFN-a-in
ducible and IFN-gamma-inducible IRF1 and IRF2. With a 6-16 ISRE probe,
therefore, M and G each correspond to two complexes which co-migrate
in band-shift assays, one corresponding to IRF1, the other to IRF2. Wi
th this probe, the constitutive complex C1/C2 corresponds predominantl
y to IRF2. Consistent with this, IRF1 and IRF2 have lower affinity for
the 9-27 ISRE than the 6-16 ISRE, whereas the reverse is true for E (
ISGF3) and its gamma subunit. Relatively small differences in affinity
appear sufficient to determine whether or not a band-shift complex is
detected. In the case of IRF1 and IRF2, the different affinities for
the 6-16 and 9-27 probes are dominated by a dinucleotide sequence in t
he centre of the 14-nucleotide 'score' ISRE. In contrast, preferential
binding of E (ISGF3) by the 39-nucleotide 9-27 ISRE-containing sequen
ce, although ISRE dependent, appears to be mediated by sequences 3' of
the 'core' ISRE. Accordingly, these complexes can be simultaneously a
ssayed using a hybrid probe consisting of the 5' flanking region and '
core' ISRE sequences from the 6 - 16 gene and sequences immediately 3'
of the 'core' 9-27 ISRE sequence. No evidence was obtained for a modu
latory role in factor binding for a pseudo-ISRE sequence close to ISRE
in the 9-27 gene. The precise roles of IRF1 and IRF2 in the induction
of IFN-beta and the control of interferon-inducible gene expression r
emain to be established. Results from the analysis of the expression o
f IRF1 and IRF2 in wild-type and mutant cells argue against a dominant
negative role for unmodified IRF2.