charges. This article must therefore be hereby marked advertise,nent 18 U.S.C. Section 1734 solely to indicate this fact.
I This work was supported by Grant CA6I862 from the NIH

in accordance with

and Special Research
quence analysis of the promoter regions of the NQO1 gene from

(c) and I 16of fac and

both mitomycin C-sensitive and -resistant HCT 116 sublines; the stability of mRNAs of NQO1 ; and (d) quantitative analysis Initiative Support from the University of Maryland School of Medicine. 2 To whom requests for reprints should be addressed. at University of Maryland Cancer characterization of the NQO1 locus in HCT 116 and HCT Center, 655 W. Baltimore St., Baltimore, MD 21201. Phone: (410) 328-3685: Fax: (410) R3OA cells. Our data demonstrate that the down-regulation 328-6559. 3 The abbreviations used are: NQO1, human NAD(P)H:quinone oxidoreductase; @-ga1, NQO1 expression in HCT 116-R3OA cells was controlled by 13-galactosidase; FISH, fluorescence in situ hybridization; FBS. fetal bovine serum; tors other than mutations to DNA sequences of the promoter RT-PCR. reverse transcription-PCR: CMV. cytomegalovirus: DCPIP, 2,4-dichlorophenol indophenol. the coding regions of the gene. 5253
MITOMYCIN C RESISTANCE AND NQO1 tN TUMOR CELLS

MATERIALS Reagents

AND METHODS
a 5% polyacrylamide gel in TBE buffer (0.045 M Tris-borate and 1.0 mist EDTA) plus 1% SDS. Gels were vacuum dried and exposed to X-ray film for
autoradiography. Radioactivity of DNA bands in gels was measured with a
Vectors (pRC/CMV and pCR II) and eDNA cycle kits were from Invitrogen Co. (San Diego, CA). The p@3-gal vectors were from Clonlech (Palo Alto, CA). STAT-60 was from Tel-Test (Friendswood, TX). Enhanced chemilumi
nescence membranes, Western blotting detection system, nucleotides Hybond-ECL nitrocellulose Corp. and 32P- and 35S-labeled were from Amersham
Betascope 603 blot analyzer (Betagen, Waltham, MA). Results of duplicate
analyses were plotted for radioactivity against the eDNA dilutions. Quantities
of NQO1 eDNA were calculated relative to the /3-actinstandard. To determine cDNAs with different sequence at nucleotide 465, unlabeled PCR products of the same NQO1 fragment were digested by AccilI. Digested products were
separated by a gel containing 3% NuSieve 3:1 and 1% agarose in TAE buffer
(Arlington Heights, IL). Reagents for PCR and the assay system of a-gal were from Promega Corp. (Madison, WI). NuSieve 3:1 agarose was from FMC
BioProducts Biochemical (Rockland, (Cleveland, ME). OH). Sequenase Prep-A-Gene II kits were from United States mem and Zeta-probe GT nylon
(0.04 M Ti-is-acetate and 1.0 mist EDTA) and analyzed by ethidium bromide

stain. Analysis of the NQO1 Gene. Genomic DNA was isolated by the method
described by Blin and Stafford (24) and digested overnight with BamHI. A
branes were from Bio-Rad (Hercules, CA). Whole chromosome libraries needed for the FISH were from Oncor (Gaithersburg, MD) and Life Technol
ogles, Inc. (Gaithersburg, MD).
target 450-bp fragment, from intron 3 to intron 4 according to the sequence of
Three clones of NQO1 eDNA (full-length clones, pDT9 and pDT2O, and exon 4 deleted clone, pDT23) from HCT 116 cells were isolated earlier (18). Rabbit polyclonal antibody was produced against recombinant NQO1of pDT9. All primers needed for PCR and DNA sequencing were based on sequences of the eDNA (19) and the gene (4) of human NQO1. Primers required for human
@-actin were based on the gene sequence reported by Ng et a!. (20).
human NQO1gene (4), was amplified from BamHI-digested DNA at dilutions between 2 and 50 ng with HQR-4043F and HQR-4472R (Table 1) as primers. The PCR mixtures and conditions were the same as described above. PCR products were separated by a 2% agarose gel and stained with ethidium
bromide determined amount tration alleles for photography. of NQOI of genomic at codon The intensities of DNA bands Dynamics, in the negatives Sunnyvale, relative were by a SC densitometer (Molecular CA). The of the NQO1 for 1 h, and
gene in each cell line was calculated DNA. To distinguish 139, amplified products different were digested

to the concen

DNA sequences by MspI

Cell Lines

HCT 116 sublines were maintained and subcultured in McCoy5A medium as described previously (17). HT-29 human colon carcinoma cells were main
tamed in the same manner as HCT cells. Breast cells lines MCF7 and
digested fragments were separated by 2% agarose gels and analyzed directly by ethidium bromide stain. Stability of mRNA HCT I 16 and HCT 116-R3OAcells were plated at 5 X 106cells per 75-cm2
culture flask and grown overnight. Cells were then treated further. with actinomycin D at 10 @zg/mlin fresh medium Northern processed blot analysis. in parallel. and incubated At time intervals of 0, 6, were
MDA-MB-23l were maintained in DMEM/Ham's F-12 medium, pH 7.4,
containing 20 mM HEPES, 5% heat-inactivated FBS and 50 @Wml entamicin g
as described by Fontana (21). Human myeloid HL6Ocells were kept in RPM!

1640 supplemented

(22). Lung
with 10% heat-inactivated

NCI-H125

FBS according to Collins et a!.

were maintained in RPM!

cell lines

and NCI-H1688

18, and 24 h, total RNA was extracted from treated cells and subjected to
As controls, cells treated with only the vehicle
1640 and 10% FBS. Monkey kidney Cos7 cells were maintained in DMEM supplemented with 10% FBS.
PCR Analysis of NQO1, Its mRNA and Gene
Analysis of mRNA. A fragment of the NQO1 Briefly, mRNA was amplified by from
Western, Northern, and Southern Analyses The procedures for Western and Northern blotting analyses were reported earlier (18). Southern blot analyses were performed according to Southern (25). Genomic DNA (10 @tg) digested overnight by PstI or Hindl!I, and was the fragments were separated by 0.8% agarose gels. DNA fragments in gels were trans-blotted to Zeta-Probe GT membranes. NQO1 eDNA of 1.3 kb was
labeled by random priming and used to detect DNA bands of the NQO1 gene.
RT-PCR for semiquantitative measurement by the method described by Hori
koshi et a!. (23) with some modifications. was performed Using HQR-298F with a eDNA cycle total RNA was isolated to suppliers'
85% confluent cells with Stat-60, and reverse transcription of mRNA to eDNA
kit according instructions. of and HQR-600R (Table 1) as primers, the target fragment
eDNA between nucleotides 298 and 617, according to the eDNA sequences of human NQO1 (19), was amplified as described earlier (18). The PCR mixture of 50 ,d contained 10 miiiTi-is(pH 8.3), 50 mMKC1,2.0 mMMgC12,0.2 m@i Cloning the Promoter
each ofdATP, dilutions. dGTP, UP, and dCTP, included: plus 10 @Ci f [32P]dCTP, o 1.0 @tM each

Regions of NQO1

@g)from each cell line was digested overnight with
of the primers, 2.5 units of Taq polymerase, and eDNA in appropriate serial
The PCR conditions 1 mm at 94C, 1 mm at 55C, 2 mm

Genomic

at 72C 30 cycles, and a final 7 mm annealing at 72C. for Simultaneously, fragments of 243 bp between exons 4 and 5 of the human (3-actin gene were
amplified primers. sample as internal standards using @A-2l05F and @3A-2424R (Table 1) as
Hindl!I and BclI and purified by phenol extraction. Nested primers (Table I), HQR-780F and HQR-2412R as the outer pair and HQR1O13Fand HQR22O1R
as the inner pair, were used to amplify 369 of the NQO1 cycling conditions gene. Contents the regions between above positions 37 and 8 of PCR mixtures (50 pA) and the thermo except for the template.

were the same as described
Radioactive PCR products in 10 @.d denatured by mixing with 1 p3 of were
buffer (0. 1 M NaOH, 10 nmi EDTA, and 1% bromphenol blue in
formamide) and heating for 2 mm at 100C. Final samples were separated with
Digested genomic DNA (1 zg) was the initial template, and products of the first PCR amplification (1 pJ) served as the second template. PCR products, showing a major DNA band of 1.2 kb by agarose gel separation, were directly
ligated to a cloning vector, pCR II. Clones were selected and expanded for
DNA sequencing. DNA sequencing by the dideoxy as described chain-termination earlier (18). method using Sequenase 2.0 was performed
Table 1 PCR primers usedfor fragmentsPurpose 3')Fragment of NQO1 eDNA Fragment of (3-actin
the ampltfication of various DNA (5' to
PrimersSequences HQR-298F'@ HQR-600R @3A-2l07F

Construction

of pRC/CMV
and NQO1 cDNA Recombinant
CAGAATC1TGGAGTfGAC CGGGAAATCGTGCGTGAC

Plasinids

NQO1 eDNA was excised by Hindlll from three clones (pDT9, pDT2O, and

eDNA Fragment of NQO1

gene NQO1 promoter (inner)

@A-2424R HQR-4043F

CACTfCATGATGGAGU
GC1TFACTCGGACCCACTC GAAGCTCCATCTCAAACAAAC
HQR-4472R HQR-1013F TACAGAC@CACC HQR-220lR CAAGGTCAAG'ITI'CTCCT NQO1 promoter (outer) HQR-780F CTAGUC!ITITCCUCACCT HQR-24l 2RCAGATATTGTGGCTGAAC CTCCACTCCTGTCCACCAATCa
F for forward and R for reverse primers.
pDT23) that were isolated earlier (18). After separation by 1% agarose gel, eDNA bands were purified by Prep-A-Gene. Purified eDNA was inserted at
the Hind!!! site of a mammalian binant plasmids, pRC/CMV-DT9 tryptophan 139 form eDNA), vector, pRC/CMV, producing three recom (wild-type eDNA), pRC/CMV-DT2O (the and pRC/CMV-DT23 (exon 4 deleted eDNA).
Each recombinant was expanded and purified. 5254
MITOMYCIN C RESISTANCE AND NQO1 IN TUMOR CELLS
Transient Transfection and Expression of NQO1 cDNA

Recombinants

Transfections scribed previously

Table 2 Semiquantitative

in Cos7 Cells
were conducted by the calcium phosphate method as de culture
analysis of NQO1 mRNAs in HCT 116 sublines by RT-PCR compared to enzymatic activity
(26). Cos7 cells were plated at 106 cells per 60-mm2
The method by Horikoshi et al. (23) was followed with modifications that were described in Materials nd Methods.Duplicate analyses were plotted for radioactivity a against the eDNA dilutions. Then calculations were normalized with an internal standard. a 243-bp fragment of )3-actin eDNA. QO1mRNA//3-actinReduction of menadione
plate. Recombinant plasmids (10 @g) p@ga1-control (5 jtg) were cotrans and fected into each plate of Cos7 cells. Transfection with the vector plasmid
pRC/CMV was carried out for basal level controls. Transfected cells were
incubated for 48 h before analysis.
deletedHCT1164.150.323.240.371536297HCT 4 Cell linesRelative (nmollmg/min)Full-lengthExon
116-R3OA0.62 0.041.49 0.1052 19

Expression

of Exon 4 Deleted NQO1 Clone in E. coli
I protein was expressed by the exon 4 deleted NQO
published earlier (18) were followed. system.

Recombinant

in E. co/i. Procedures

The activity

of a-gal
was used for the normalization

of transfection

Preparation of Cell Extracts and Enzyme Assays
Methods published previously (17) were used to prepare soluble extracts
from HCT the plate I 16 and HCT by the lysis 1 !6-R3OA buffer cells and to assay NQO1 by the reduction /3-gal assay system at 12,000 according to the

efficiency.

Cytogenetic Studies
Cytogenetic analysis was performed on HCT cells on the third passage number after its
of menadione-cytochrome c or DCPIP. Transfected cells were lysed directly in
in the suppliers' activity instructions. studies. Cell lysates of were centrifuged X g for 10 mm of
after it was obtained from American Type Culture Collection (passage unknown) and on the HCT 1 16-R3OA cell line on the fourth passage isolation. examined molecular Cells were cultured at the studies 500-band with FISH, labeled one instructions. round by standard level slides techniques. Metaphase karyotype
to obtain cell extracts that were used for Western blotting and enzymatic
The activity @-ga1was determined by the production
preparations analysis. For

were Giemsa-banded

using trypsin (27). Twenty cells from each cell line were

for conventional containing metaphase cells were hybrid green, according possi cells were counter

o-nitrophenol

from o-nitrophenyl-@3-D-galactopyranoside

using the a-gal assay

ized with probes to the suppliers' ble. Following stained
with digoxigenin, Dual-color of signal
biotin, or spectrum hybridization
was used when the slides metaphase
amplification, microscope

and viewed

in a fluorescence

(28). Twenty

from each cell line were examined.
RESULTS Measurement of NQO1 mRNAs. DNA fragments of two different sizes were generated by RT-PCR amplification with primers HQR 298F and HQR-600R from NQO1 mRNA in both HCT 116 and HCT 1!6-R3OA cells (Fig. 1). These fragments, 320 and 206 bp, repre sented mRNA coding full-length cDNA and exon 4 deleted cDNA, respectively. The amount of mRNA coding full-length cDNA relative to @-actin HCT 116-R3OA cells was only 15% of that present in in HCT 116 cells (Table 2). In HCT 116-R3OA cells, the amount of exon
4 deletedmRNA wastwice thatof full-length but only one-halfof the
amount present in HCT 116 cells. The enzymatic activity of NQO1 corresponded to the amount of the full-length mRNA but not the total mRNA. The 206-bp DNA band was also found in RT-PCR-amplified products of seven other human tumor cell lines, including colon cancers HT-29 and BE, breast cancers MCF-7 and MDA-MB-23l, lung cancers NCI-Hl25 and NC!-Hl688, and myeloid leukemia HL-60 (Fig. 1). The percentage of the exon 4 deleted mRNA in each of the seven cell lines was substantially less than the full-length mRNA. The heterozygosity of nucleotide 465 in full-length cDNAs was analyzed by the loss of a restriction site of AccI!!. The 320-bp DNA of the wild type (arginine 139, TCCGGA) yields two fragments of 166 and 154 bp from Acc!I! digestion, whereas the tryptophan 139 form (TCTGGA) remains uncut. As we showed earlier (18), both forms of full-length mRNAs of NQO1 were detected in HCT 116 cells; but HCT 1l6-R3OA cells expressed only the tryptophan 139 form. Among the seven other cell lines tested, only HL-60 expressed both forms of NQO1 mRNA, whereas the other six cell lines expressed solely the wild type. Northern blot analysis of total RNA showed that two species of NQO1 mRNAs (I.2 and 2.7 kb from different polyadenylation sites) were present in both HCT 116 and HCT 1l6-R3OA cells (data not shown). Each mRNA species in the two cell lines decayed at a similar rate of I2% 3% per 10 h for each species in both cell lines.

206 166

UA U A U A U
Fig. I. Agarose gel separation of PCR-amplified fragments of NQOI eDNA from various cell lines. RT-PCR amplification of an mRNA fragment between 298 and 617
were conducted as described in the text. PCR products obtained from each cell line were
purified by ethanol precipitation. One-half of the purified PCR product was digested by AcdllI (A) and one-half remained intact (U). The prepared samples were separated by a gel containing 1% agarose and 3% NuSieve 3: 1 agarose and analyzed. 5255
Expression of Exon 4 Deleted cDNA of NQO1 in E. coil and
Cos7. A protein of a molecular size of Mr 26,000 was expressed in E. co/i by the exon 4 deleted NQO1 cDNA clone (pDT23). Western blot analysis showed that this protein interacted with polyclonal antibody against recombinant NQO1 (anti-NQO1) equally well as NQO1 pro duced by HCT 116 cells (data not shown). However, it failed to reduce common substrates of I such as menadione and DCPIP, with either NADH or NADPH as electron donors. Western blot analysis did not detect a Mr 26,000 protein corre sponding to the translation of the exon 4 deleted mRNA in extracts of any of the cell lines. Recombinant plasmids of exon 4 deleted cDNA (pRC/CMV-DT23) failed to express a Mr 26,000 protein that would interact with anti-NQO1 in Cos7 cells (Fig. 2). However, similar recombinant plasmids of full-length cDNA, pRC/CMV-DT9 (wild type), and pRC/CMV-DT2O (tryptophan I39 form) expressed a Mr 31,000 protein over and above the basal level in Cos7 cells. Proteins produced by both full-length recombinants showed full enzymatic activities (Table 3). The constructs of a rat NQO1 cDNA also ex pressed a Mr 31,000 protein in Cos7 cells but showed higher enzy matic activity than human clones.

1358 815

@,. *1 @

267: 174

434/@i
NQO1 Gene in HCT 116 Sublines. To evaluatethe genetic source
of the two full-length NQO1 mRNAs coding the arginine form (wild
Fig. 3. Agarose gel separation of PCR amplified fragments of the NQOJ gene from HCT I 16 sublines. A 450-bp fragment of the NQO1 gene from each cell line was amplified by PCR as described in the text. Portions of the PCR products were digested by MspI. Intact (U) and Mspl-digested (M) PCR products of equal amount were separated by a 2% agarose gel. Gels were examined after ethidium bromide stain.
Table 4 Semiquantitative analysis of a fragment of the NQO1 gene in HCT I 16sublines
PCRA j@ 45O-bp fragment of the NQO1 gene between intron 3 and intron 4 was PCR weredigested amplified as described in Materialsand Methods. Portions of PCR products wereseparated MspI. PCR products, digested and intact controls, of equal amount by bydensitometer.a 2% agarose gel. The intensity of each DNA band was determined by concentrationof The amount of the NQO1 gene was calculated relative to the genomic DNA.Relative geneCell formHCT lines 1.17HCT 116.97a 116-R3OA

ND, not detected.

intensity of NQO1 Wild type 0.98 ND Tryptophan I

I3lkDa

Fig. 2. Westem blot of NQO1 expressed by Cos7 cells after transient transfection of eDNA recombinant plasmids of NQO1 eDNA. Methods including the construction of recombinant plasmids, transfection, and preparation of cell extracts were described in the text. Proteins (20 @.sg) from each crude cell extracts were separated by a SDS-polyacryl amide gel (12.5%). blotted to nitrocellulose membranes, and detected by enhanced chemiluminescence system as described earlier (18). Recombinant plasmids used to transfect Cosl cells: Lane 1, pRC/CMV-DT9 (wild type eDNA); Lane 2, pRC/CMV (vector); Lane 3. pRC/CMV-DT23 (exon 4 deleted eDNA); and Lane 4. rat NQO1 eDNA.
Table 3 Transient expression of NQO1 in Cos7 cells by recombinant plasmids ofNQO1
liverMethods cDNA obtainedfrom HCT I 16 cells and rat preparationof including the construction of recombinant plasmids. transfection, SEfor extracts and enzyme assay are described in the text. The data represent mean cell at least three experiments of duplicate analyses.NQO1 eDNA recombinants DCPIPin (nmol/mg/min)None cells Cos7 645pRC/CMV 15pRC/CMV-DT9 52pRC/CMV-DT2O (wild type) 132pRC/CMV-DT23 (tryptophan 139) 18pRC/CMV-RDT (Exon 4 deleted) (rat NQO1 eDNA) Reduction of
type) and the tryptophan 139 form, we amplified a 450-bp fragment of the NQO1 gene from intron 3 to intron 4 containing exon 4 and codon 139. The conversion from wild-type sequence (CGG) to the trypto phan 139 form (CTGG) destroys a restriction site of MspI. A 450-bp band was revealed by gel electrophoresis from the PCR product obtained from the NQO1 gene in HCT 116 cells (Fig. 3). One-half of this 450-bp band was digested by Msp! into two fragments of 100 and 350 bp, and one-half of the band remained unchanged. A 450-bp band was also produced from the NQO1 gene in HCT 1l6-R3OA cells, but the entire band was not digested by MspI. Semiquantitative PCR analysis showed that relative amounts of the NQO1 gene in the two cell lines were similar (Table 4). However, two forms of the gene were present in relatively equal proportion in HCT I 16 cells; but only the tryptophan 139 form was present in HCT 1l6-R3OA cells. Southern blot analysis of the NQO1 gene by hybridization with a NQO1 cDNA probe (Fig. 4) showed that the size and the intensity of DNA bands of the two cell lines were similar.
Promoter Regions of NQO1. A l206-bp fragmentof the NQOI

1263 181

gene, including 837and 369 bp from the startpoint, was cloned from both HCT 116 and HCT I l6-R3OA cells. Five clones from each cell line were isolated and sequenced. All clones from both sources showed similar DNA sequences, corresponding to the same region of the human NQO1 gene sequence (4). 5256

.-. -@.

abnormalities as the parent HCT 116 cells. In addition, extra material was found on one chromosome 5 in HCT 1l6-R3OA cells, and its source was not identified.

--@@-21 8

DISCUSSION Various studies have shown an association of mitomycin C resist ance and a deficiency of NQO1 in tumor cell lines (1 118). However, this association remains controversial. An association of NQO1 activ ity and aerobic mitomycin C sensitivity has been demonstrated in several studies. These included the correlation between mitomycin C sensitivity and NQO1 activity in tumor xenografts reported by Malkinson et a!. (29). More convincingly, the correlation was exhib ited by Fitzsimmons et a!. (30) with a panel of 69 NIH human tumor cell lines. On the other hand, a lack of correlation of NQO1 activity and cytotoxicity of mitomycin C was reported in a panel of 15 cell
lines by Robertson et a!. (31). Furthermore, Nishiyama et a!. (32)

@ 3.4 -

2.3 2.0

PstI Hindlil

Fig. 4. Southern blot of the NQO1 gene in HCT 116 sublines. Genomic DNA (10 zg) was digested overnight by Hindill and PstI, and the fragments were separated with a 0.8%
agarose gel. Southern blotting was performed according to published procedures (26). A
1.3-kb eDNA probe was labeled by random priming and used to detect DNA bands of the NQO1 gene. Lanes 1 and 3, HCT 116 cells; Lanes 2 and 4, HCT 116-R3OA cells.
reported that high sensitivity to mitomycin C correlated with high NQO1 activity in cultured tumor cells; but the relationship was the opposite in tumor xenografts. Despite the controversy, the importance of NQO1 in human tumor cells and human tissue applies not only to the activation of mitomycin C to cytotoxic intermediates but also to the activation and detoxification of other bioreductive drugs such as 3-hydroxy-5-aziridinyl-1-methyl-2-(1H-indole-4,7-indione)-prop-f3en-a-I, dimethyldiaziridinyl benzoquinone, and tirapazamine. The
activities of NQO1 and other reductive enzymes and their optimal
Cytogenetic Analysis. G-banded karyotyping showed that all HCT 116 cells had three abnormal chromosomes with extra material (Fig. 5). The Y chromosome was missing from many cells. FISH, with labeled chromosome libraries, was used to identify the source of these extra pieces (data not shown). Chromosome 10 had a partial duplica tion at the distal tip of the long arm. There were two unbalanced translocations; an extra piece of the chromosome 8 long arm was attached to the short arm of chromosome 16, and an extra piece of the chromosome 17 long arm was attached to the short arm of chromo some 18. Thus, there was partial trisomy for chromosomes 8, 10, and 17. FISH, with a chromosome 16 library (data not shown), did not reveal any chromosome I 6 material translocated to any other chro mosome. As for HCT 116-R3OA, all cells had the same pattern of physiological conditions all play a role in the cytotoxicity of mito mycin C and other bioreductive antitumor agents. A recent review by Workman (3) discussed these subjects in detail. The cause of NQO1 deficiency in our mitomycin C-resistant sub line, HCT 116-R3OA, can be explained by the decreased amount of full-length mRNAs, the mRNA responsible in producing active en zymes. Furthermore, the resistant cells did not have the wild-type mRNA that was present in the parent HCT I 16 cells. Analysis of the NQO1 gene in HCT 116-R3OA cells demonstrated that the lack of wild-type mRNA was due to the loss of heterozygosity at the NQO1 locus. The concentration of the NQO1 gene measured by both PCR amplification and Southern blot showed that the loss of heterozygosity and the low expression of full-length mRNA in HCT 116-R3OA cells

Fig. 5. Giemsa-banded

partial karyotypes
HCT 116 (top panel) and HCT 116-R3OAcells
(bottom panel). Abnormal chromosomes, 10, 16 and 18, from both cell lines and chromosome 5 from HO' 116-R3OA are presented. In each chro mosome pair, the abnormal chromosome is to the right of its normal homologue.

.11 1,

MITOMYCIN C RESISTANCE AND NQO, IN TUMOR CELLS
did not result from a loss of copies of the gene. Karyotypic analysis confirmed these results. In searching for the mechanisms that regulated the expression of NQO1 in HCT 116-R3OA cells, we have ruled out two possibilities. One was the stability of the tryptophan 139 form of mRNA, which did not decay faster than the wild-type mRNA. The other possibility was mutations to DNA sequences of the promoter regions of the NQO1 gene (837 bp upstream) in HCT 116-R3OA cells. No mutations were seen within this 837-bp region where several regulating cis elements of the NQO1 gene lie (46).However, we are concerned about the possibility of mutations to unknown regulatory elements that are located beyond the 837-bp region. We do not believe that the substi tution of amino acid 139 from arginine to tryptophan has any effect on the expression of QO1 because plasmid constructs of these two types of cDNA and a CMV promoter expressed NQO1 equally well in HCT 1l6-R3OA cells.4 Exon 4 deleted mRNA, a result of alternative splicing, was detected in a wide spectrum of cell lines and normal tissues by Gasdaska et a!. (33) and in our present study. The function of exon 4 deleted mRNA is yet to be found. At present, the relation ship between the loss of the wild-type allele of the NQOJ locus in the mitomycin C-resistant cells and the low expression of QO1 in these cells is under investigation. Factors such as trans-acting factors re ported by Jaiswal (4), Li and Jaiswal (5), Xie et a!. (6), and Yao et a!. (34, 35) and alternative splicing of the mRNA of NQO1 probably all play a role in the regulation of the concentration of active QO1 in HCT 116-R3OA cells. The location of the human NQO1 gene was reported to be on the long arm of chromosome 16 (19). Our study showed that the loss of homozygosity of NQO1 locus in HCT 116-R3OA cells did not result from a loss ofone chromosome 16 or 16q or translocation of 16q. This result is in contrast to an earlier report by Wilson et al. (36), who showed that a copy of chromosome 16 was lost in their mitomycin C-resistant subline, HCT 116R22. Although both mitomycin C-resist ant cell lines from the two laboratories originated from the same parent cell line, HCT 116, it is likely that different mitomycin Cresistant clones were selected in the process. The abnormality to one copy of l6p was observed in both HCT 116 and HCT 116-R3OA cells, but it was not seen in either the parent or the mitomycin C-resistant cells in the earlier report (36). This may indicate karyotype evolution in cultures. Our current data suggested that HCT 1l6-R3OA probably was not a mutated daughter cell clone of HCT 116 due to mitomycin C treatment. Most likely, it was a preexisting cell clone that did not have

gene. There remains a great deal more to be learned about the regulation of NQO1 in all cells. ACKNOWLEDGMENTS We are grateful to Drs. Nicholas Bachur and Merrill J. Egorin for critically
reviewing the manuscript.

REFERENCES

1. Ernster, L., Danielson, L., and Ljunggren, M. DT-diaphorase. 1. Purification from the soluble fraction of rat-liver cytoplasm and properties. Biochem. Biophys. Ada, 58: 171188, 1961. 2. Riley, R. J., and Workman, P. DT-diaphorase and cancer chemotherapy. Biochem. Pharmacol., 43: 16571669,1992. 3. Workman, P. Enzyme-directed bioreductive drug development revisited: a commen tao' on recent progress and future prospects with emphasis on quinone anticancer agents and quinone metabolizing enzymes, particularly DT-diaphorase. Oncology Res., 6: 461475, 1994. 4. Jaiswal, A. K. Human NAD(P)H:quinone oxidoreductase (NQO1) gene structure and induction by dioxin. Biochemistry, 30: 1064710653. 1991. 5. Li, Y., and Jaiswal, A. K. Regulation of human NAD(P)H:quinone oxidoreductase gene: role of API binding site contained within human antioxidant response element. J. Biol. Chem., 267: 1509715104, 1992. 6. Xie, T., Belinsky, M., Xu, Y., and Jaiswal, A. K. ARE- and TRE-mediated regulation of gene expression: response to xenobiotics and antioxidants. J. Biol. Chem., 270: 68946900, 1995. 7. Schlager, J. J., and Powis, G. Cytosolic NAD(P)H:(quinone-acceptor)oxidoreductase in human normal and tumor tissue: effects of cigarette smoking and alcohol. Int. J. Cancer, 45: 403-409, 1990. 8. Cresteil, T., and Jaiswal, A. K. High levels of expression of the NAD(P)H:quinone oxidoreductase (NQO1) gene in tumor cells compared to normal cells of the same origin. Biochem. Pharmacol., 42: 10211027,1991. 9. Schor, N. A., and Cornelisse, C. J. Biochemical and quantitative histochemical study of reduced pyridine nucleotide dehydrogenation by human colon carcinomas. Cancer Res., 43: 48504855, 1983. 10. Koudstaal, J., Makkink. B., and Overdiep, S. H. Enzyme histochemical pattern in human tumors. II. Oxidoreductases in carcinoma of the colon and the breast. Eur. J. Cancer, 11: 111115, 1975. 11. Traver, R. D., Horikoshi, T., Danenberg, K. D., Stadlbauer, T. H., Danenber, P. V. Ross, D., and Gilbson, N. W. NAD(P)H:quinone oxidoreductase gene expression in human colon carcinoma cells: characterization of a mutation which modulates DT diaphorase activity and mitomycin sensitivity. Cancer Res., 52: 797802, 1992. 12. Eickelmann, P., Schulz, W. A., Roche, D., Schmitz-Drager, B., and Sies, H. Loss of heterozygosity at the NAD(P)H:quinone oxidoreductase locus associated with in creased resistance against mitomycin C in a human bladder carcinoma cell line. Biol. Chem. Hoppe-Seyler. 375: 439445, 1994. 13. Marshall, R. S. Paterson, M. C., and Rauth, A. M. Deficient activation by a human cell strain leads to mitomycin resistance under aerobic but not hypoxic conditions. Br. J. Cancer, 59: 341346,1989. 14. Marshall, R. S., Paterson, M. C., and Rauth, A. M. Studies on the mechanism of resistance to mitomycin C and porfiromycin in a human cell strain derived from a cancer-prone individual. Biochem. Pharmacol. 41: 13511360,1991. 15. Dulhanty, A. M., and Whitmore, G. F. Chinese hamster ovary cell lines resistant to mitomycin C under aerobic but not hypoxic conditions are deficient in DT-diapho

rase. Cancer Res., 51: 18601865, 1991.
16. Begleiter, A., Robotham, E., Lacey, G., and Leith, M. K. Increased sensitivity of quinone resistant cells to mitomycin C. Cancer Lett., 45: 173176,1989. 17. Pan, S., Akman, S. A., Forrest, G. L., Hipsher, C., and Johnson, R. The role of NAD(P)H:quinone oxidoreductase in mitomycin C- and porfiromycin-resistant HCT I 16 human colon-cancer cells. Cancer Chemother., 31: 2331,1992. 18. Pan, S., Forrest, G. L., Akman, S. A., and Hu, L-T. NAD(P)H:quinone oxidoreductase
functional NQO1 alleles (wild type) and that was isolated during the
process of selection of mitomycin C resistance. Other mitomycin C-resistant cell lines containing only null homozygous NQO1 include the BE and H596 cells described by Traver et a!. (11, 37) and the subline (RT1 12MMC) of a human bladder cancer line (RT1 12) de scribed by Eickelmann et a!. (12). In addition, Traver et a!. (38)
(NQO,) expression and mitomycin C resistance developed by human colon cancer HCT 116 cells. Cancer Res., 55: 330335, 1995.
19. Jaiswal, A. K., McBride, 0. W., Adesnik, M., and Nebert D. W. Human dioxin inducible cytosolic NAD(P)H:menadione oxidoreductase: eDNA sequence and local ization of gene to chromosome 16. J. Biol. Chem., 263: 1357213578, 1988. 20. Ng, S-Y., Gunning, P., Eddy, R., Ponte, P., Leavitt, J., Shows, T., and Kekes, L. Evolution of the functional human /3actin ene and its multi-pseudogene family: g conservation of noncoding regions and chromosomal dispersion of pseudogenes. Mol. Cell. Biol., 5: 27202732, 1985. 21. Fontana, J. A. Interaction of retinoids and tamoxifen on the inhibition of human mammary carcinoma cell proliferation. Exp. Cell Biol., 55: 136144,1987. 22. Collins, S., Gallo, R. C., and Gallagher, R. E. Continuous growth and differentiation of human myeloid leukaemic cells in suspension culture. Nature (Lond.), 270: 347349,1977. 23. Horikoshi, T., Danenberg, K. D., Stadlbauer, T. H. W., Volkenandt, M., Luke, C. C., Shea, L. C., Aigner, K., Gustavsson, B., Leichman, L., Frosomg, R., Ray. M., Gibson, N. W., Spears, C. P., and Danenberg, P. Quantitation of thymidylate synthase. dihydrofolate reductase, and DT-diaphorase gene expression in human tumors using the polymerase chain reaction. Cancer Res., 52: 1081 6, 1992. 1 24. Blin, N., and Stafford, D. W. A general method for isolation of high molecular weight DNA from eukaryotes. Nucleic Acids Res., 3: 23032308, 1976.

reported recently that 6% incidence of individuals (among 48 matched
sets of tumors and normal lung tissue) and 18% incidence of healthy Chinese from Shanghai had homozygous NQO1 with the same mu tation as the BE cell. HCT l16-R3OA cells do not have the same point mutation to its NQO1 gene as that present in the other three cell lines. However, they all share the characteristics of down-regulation of the NQO1 gene and the loss of the wild-type NQO1 allele. We feel that the down-regulation of the NQO1 gene associated with the loss of func tional alleles (wild-type alleles) in mitomycin C-resistant cells may involve other factors, such as trans-acting factors, and DNA methyl ation in addition to the mutations of the coding regions of the NQO1
4 S-s. Pan, unpublished data.
25. Southern, E. M. Detection of specific sequences among DNA fragments separated by
26. gel electrophoresis. J. Mol. Biol., 98: 503517,1975. Ausubel, F. M., Brent, R., Kingston, R. E., Moore, D. D., Seidman, J. G., Smith, J. A., and Struhl, K. Transfection of DNA into eukaryotic cells. In: Current Protocols in Molecular Biology, Suppl. 14, pp. 9.1.19.1.9.New York: Wiley and Sons, 1990. Verma, R. S., and Babu, A. Human Chromosomes: Manual of Basic Techniques. pp. 4108.New York: Pergamon Press. 1989. Lichter, P., and Cremer, T. Chromosome analysis by non-isotopic in situ hybridiza tion. In: D. E. Rooney and B. H. Czepulkowski (eds.), Human Cytogenetics, Vol. I, pp. 157192.New York: Oxford University Press, 1992. Malkinson, A. M., Siegel, D., Forrest, G. L., Gazdar, A. F., Oie, H. K., Chan, D. C., Bunn, P. A., Mabry, M., Dykes, D. J., Harrison, S. D. Jr., and Ross, D. Elevated DT-diaphorase activity and messenger RNA content in human non-small long carci noma: relationship to the response of lung tumor xenografts to mitomycin C. Cancer Res., 52: 47524757, 1992. Fitzsimmons, S., Workman, P., Grever, M., Paull, K., Camalier, R., and Lewis, A. D. Reductase enzyme expression across the National Cancer Institute tumor cell line panel: correlation with sensitivity to mitomycin C and EO9. J. Natl. Cancer Inst., 88: 259269, 1996. Robertson, N., Stratford, I. J., Houlbrook, S., Carmichael, J., and Adams, G. E. The sensitivity of human tumour cells to quinone bioreductive drugs. What role for DT-diaphorase? Biochem. Pharmacol., 44: 409412, 1992.
32. Nishiyama, M., Sack), S., Aogi, K.,, Hirabayashi, N., and Toge, T. Relevance of DT-diaphorase activity to mitomycin C (MMC) efficacy on human cancer cells: differences in s'itro and in @ivo systems. Int. J. Cancer, 53: 10131016,1993. 33. Gasdaska, P. Y., Fisher, H., and Powis, G. An altematively spliced form of NQOI (DT-diaphorase) messenger RNA lacking the putative quinone substrate binding site

27. 28.

ispresent inhumanormal n andtumor tissues. Cancer 55:25422547, Res. 1995.
34. Yao, K., Xanthoudadis, S., Curran, T., and O'Dwyer. P. J. Activation of AP-l and of a nuclear redox factor, Ref- 1, in the response of HT29 colon cancer cells to hypoxia. Mol. Cell. Biol., 14: 59976003, 1994. 35. Yao, K., and O'Dwyer, P. J. Involvement of NF-KB in the induction of NAD(P)H: quinone oxidoreductase (DT-diaphorase) by hypoxia, oltipraz and mitomycin C. Biochem. Pharmacol., 49: 275-282, 1995. 36. Willson, J. K. V., Long, B. H., Marks, M. E., Brattain, D. E. Willey, J. E., and Brattain, M. G. Mitomycin C resistance in a human colon carcinoma cell line associated with cell surface protein alterations. Cancer Res. 44: 58805885. 1984. 37. Traver, R. D., Phillips, R. M., Gibson, N. W., and Ross, D. A point mutation in both human lung and colon carcinoma cell lines leading to a loss of DT-diaphorase activity. Proc. Am. Assoc. Cancer Res., 36: 525, 1995. 38. Traver, R. D., Rothman, N., Smith, M. T., Yin, S. Y., Hayes, R. B., Li, G. L., Franklin, W. F., and Ross, D. Incidence of a polymorphism in NAD(P)H:quinone oxidoreductase (NQO1). Proc. Am. Assoc. Cancer Res., 37: 278, 1996.

responsible for the disparity of gene expression of NQO1 and the low concentration of NQO1 protein in MMC-resistant sublines. Reversal of MMC resistance and the recovery of NQO1 in two revertants further
supports the hypothesis that cellular control of NQO1 can modulate the

cytotoxicity of MMC.

INTRODUCTION In recent years, several resistance mechanisms have been observed in tumor cells that have developed resistance to MMC,3 an important agent for the treatment of solid tumors (1). One type of MMC resistance, which has been observed in an L1210 leukemia subline by DolT et a!. (2) and in a P388 leukemia subline by Rose et a!. (3), has the characteristics of multiple drug resistance. These cells exhibited cross-resistance to anthracyclines and Vinca alkaloids, and expression of membrane p-glycoprotein was detected in the L1210 cells. Another mechanism of MMC resistance involves enhancement of MMC de toxification. One such detoxification mechanism is the increase of glutathione S-transferase concentrations in HCT 116 sublines (4). Additionally, a number of MMC-resistant cell lines exhibit cellular
Received 6/23/94; accepted 11/9/94.
The costs of publication of this article were defrayed in part by the payment of page
charges. This article must therefore be hereby marked advertisement in accordance with
18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by Grant CH-412 from the American Cancer Society

(to S. P.)

2 To whom requests for reprints should be addressed, at University of Maryland Cancer
Center, 655 West Baltimore Street, Baltimore, MD 21201.
3 The abbreviations used are: MMC, mitomycin C; NQO,, human NAD(P)H:quinone
oxidoreductase; DCPIP, 2,4-dichlorophenol indophenol; nt, nucleotide; TBST, 100 mai
Tris, 0.9% NaG, 0.1% Tween 20 at pH 7.5.
QUINONEREDUCFASEAND MITOMYCINC RESI5TANCE

MATERIALS

AND METHODS
Reagents. MMC was kindly supplied by Dr. Stephen Carter, Bristol Re
search Laboratories (Wallinglord, CT). Cibacron blue 3GA, ampicillin, lysozyme, and horse anti-rabbit immunoglobulin antibodies conjugated with

alkaline phosphatase were from Sigma Chemical Co. (St. Louis, MO). DEAE
Sephacel 200, PhastGels, and silver stain were from Pharmacia (Piscataway, NJ). Enhanced chemiluminescence Western blotting detection system, Hy
of dATP, dCTP, dGTP, and dTFP, 2.5 units of AmpliTaq polymerase, 1.0 ng cDNA, and 1.0 @.tM of the two DNA primers in 100 gil Briefly, the each thermal cycling included 1-mm denaturation at 96C, 30-s annealing at 52C, and 1-mm extension at 72C 35 cycles, a fmal annealing of 7 mm at 72C, for and cool down to 4C.After analysis of the PCR products by agarose
gel-electrophoresis, a band at about 1.3 kilobase was excised from the gel and
bond-ECL nitrocellulose membranes, and [a-32P]dCTPwere from Amersham Corp. (ArlingtonHeights, IL). Prep-A-geneand Zeta-probeGT nylon mem
branes were from Bio-Rad (Richmond, CA). Other reagents were: [a-35S1
purified by Prep-A-gene according to the manufacturer's instructions. Purified PCR productwas inserted at the EcoRV site of the pBluescriptII KS(+). Recombinants were transformed into E. co/i DH5a. Colonies were selected for

ampicillin resistance.

dATP from New England Nuclear (Boston, MA), RNAzol B from Cinna
Biotech (Friendswood, TX), Copy Kit from Invitrogen Co. (San Diego, CA), GeneAmp DNA PCR kit from Perkin Elmer/Cetus (Norwalk, Cr), Sequenase 2.0 from United States Biochemical Corp. (Cleveland, OH), NuSieve 3:1 agarose from FMC BioProducts (Rockland, ME), and Random priming system
DNA Sequencing. SelectedE. coli DH5a colonies containingrecombinant

plasmids

further
were grown in Luria-Bertani
purified as described (19).

medium

(LB broth) containing

DNA sequencing

mg/liter ampicillin. Plasmids of each clone were isolated by alkaline lysis and

Double-stranded

I from New EnglandBiolabs (Beverly, MA).
Rabbit polyclonal antibody was produced against purified NQO1 that was
cloned from HCT 116 cells and expressed in E. coli JM1O9(NQO1wEcolt).All primers required for PCR and DNA sequencing were synthesized based on the cDNA sequence data and nt numbers of human liver NQO1reported by Jaiswal
et al. (15) according to the rules described by Lathe (16).
performed according to a modified Sanger (20, 21) procedure with Sequenase 2.0 and [a-35S]dATP labeling. Both DNA strands of each selected plasmid were sequenced. Nine primers, including T3 and T7 primers, three forward, and four inverse primers (Table 1) with selected sequences taken from the DNA

sequence data of human liver NQOI (15), were used.
Expression of NQO, in E. coli. DNA inserts of NQOI clones were excised
from selected Bluescript recombinants by NcoI and Hindu! digestion. The
1.3-kilobase inserts were purified with Prep-A-Gene and inserted between the
Cell Lines and Cell Survival. HO' 116 humancolon carcinomacells were
obtained from the American Type Culture Collection (Rockville, MD). The
NcoI and Hindu! sites of the expression vector pKK233.2. The expression
recombinant was transformed into E. coli strain JM1O9. Colonies with positive
development of the MMC-resistant HCI' 116-R3OA subline was described
previously (12). Other MMC-resistant sublines, designated as RiO, R25, and R40, were isolated after exposure to MMC at increasing concentrations up to
expression of NQO1 activity were selected.
Isolation of NQO1 PrOtein Expressed by E. coli JM1O9. E. coli JM1O9 colonies with positive expression of NQOI were grown in LB broth containing
10, 25, and 40 ,tM, respectively. All HCT 116 sublines were maintained and 75 mg/liter ampicillin to an absorbanceof 1.21.5. Cells were recovered, subcultured as described previously (12). MMC resistance of HCT 116-R3OA washed with cold 50 mMTris-HC1at pH 7.4, and then resuspended at 0.1 g/ml
cells reverted when subcultures were maintained free of MMC for more than
1 year. Two sublines were isolated on separate occasions when MMC resist
ance reverted and were designated HO' 116-R3OArtl and HO.' 116-R3OArt2.
The cytotoxicity of MMC to HCT 116 sublines was assessed by a modified colony formation assay described by Rockwell and KalIman (17). Details of
the procedure have been described previously (12). The drug concentration
producing 50% inhibition of cell growth for each cell line was calculated by
median-effect analysis with a program developed by Chou and Chou (18). Cloning of NQO1 cDNA. Total RNA was isolated from 85% confluent
in buffer A containing 50 mM Tris, 0.5 mM each of DIT, EDTA, phenyl methylsulfonyl fluoride, and benzoamidine at pH 7.4. After the addition of lysozyme at 50 p@g/ml,the cell suspension was incubated at room temperature for 15 mm and sonicated with a Branson 250 Sonifier (Danbury, C'F) for 10 cycles in 10-s pulsing and cooling on ice for every 100 ml of cells. Complete homogenization was confirmed by microscopic observation. Sucrose and streptomycin-sulfate were added to the crude homogenate with stirring to 0.25

freeze-thaw method as described previously (12). Total soluble fractions were

used in SDS-PAGE

with a 9% resolving

NQO,QR-600Forward

NQO,QR-315Inverse
CAGATAUGTGGCITIAAC nt 298-315 of GTCVfAGAACCTCAACfG

nt 600-617 of

for Western blouing as described below.
gel to separate proteins. Unstained gels were trans-blotted
Western Blotting. SDS-PAGE adapted from Laemmli (23) was performed
to Hybond-ECL membranes nitrocellulose membranes with 50 mM Tris, 40 mM glycine at
NQO,QR-617Inverse NQO1QR-923Inverse
GTTCAGCCACAATATCFG nt 315-298 of CAGUGAGG11@CTAAGAC nt 617-600 of ATACCCAGA1TFGATAAC nt 923-906 of NQO,
pH 9.2, and 20% methanol by a Bio-Rad Semi-Dry Transfer cell. Blotted
were blocked with 5% nonfat dry milk in TBST, washed with chemiluminescence Western blotting detection system.
TBST, and blotted dry. Protein bands of NQO1on the membrane were detected

by an enhanced

a Selected sequences of all primers were taken from the DNA sequence data of human liver NQO, cDNA reported by Jaiswal et aL (15) according to the rules described by Lathe (16). 331
Briefly, the blots were incubated first with rabbit anti-NQO1w'@ antibody, COil followed by blocking with 5% nonfat dry milk in TBST and 0.1% Tween 20,
and rinsing with TBST. Then the membrane was incubated with antirabbit IgG
QUINONEREDUCMSE AND MITOMYCIN RESISTANCE C
antibody conjugated to horseradish peroxidase. Antigen-antibody reacting
bands were detected by the oxidation of luminol by peroxidase in the presence
of hydrogen peroxide and enhancer. Light emission was captured by exposing
blots to KOdak XAR-5 films which were scanned with a Beckman DU7O spectrophotometer. Based on equal amounts of total protein, the intensity of
NQO1 bands for each subline was calculated as a percentage of the same

protein in HCT 116.

Northern Blot Analysis. To probe NQO1messengers, an NcoI and Hind!!!
cut fragment (1.3 kilobase) of a cDNA clone was labeled with [a-32PIdCTP to
a specific activity of 0.8 X 10 cpm/g@g random priming with a New by England Biolabs kit. Labeled probes were purified by Sephadex G-25 Spin columns. Northern blot of total RNA was conducted as described previously
(12). Intensity of radioactivity for each mRNA band on Northern blots was measured with a Betascope 603 blot analyzer (Betagen, Waltham, MA). In
order to account for variability in gel loading and transfer efficiency, mem
cell growth values under aerobic conditions of each subline were determined (Table 2). HCF 116-R25 and HCF 116-R40 were as resistant as HCT 116-R3OA, and HCT 116-RiO was somewhat less resistant than HCT 116-R3OA. Sensitivity to MMC reverted to paren tal level in HCF 116-R3OArtl and HC1' 116-R3OArt2. NQO1 Clones. PCR amplification of NQO1 cDNA templates of both HCT 116 and HCT 116-R3OA cell lines produced a DNA band of about 1.3 kilobase by agarose gel-electrophoresis analysis. After ligating the PCR products to Bluescript II KS and transforming recombinants into E. coli DH5a, we selected eight clones: three clones from HCF 116 and five from HCT 116-R3OA.

DNA Sequences. Double-strand DNA sequencing of all isolates
showed that inserts of two different lengths were cloned from each cell origin. One clone each from HCF 116 and HCF 116-R3OA, designated as pDT9 and pDT2O, respectively, had the full length of the designed DNA sequence (nt 91313of human liver NQO1 cDNA) including the entire coding region. The first 867 base pair of pDT9 matched the sequence nt 9875of the reported human liver NQO1 cDNA (15) with the exception of two base pairs. This included 42 base pairs upstream of the initiation codon and the entire coding region. The two changed base pairs were substitutions at nt 98 from
branes were stripped and reprobed with a fragment of human f3-actin cDNA (2
kilobases). The radioactivity of (3-actinmeasured by Betascope was used to
normalize the radioactive counts of NQO1 mRNA bands for each cell line.
Levels of each species of mRNA of N001 in all of the sublines were then
calculated as the percentage of the same mRNA species in HCF 116 cells. Reverse Transcription-PCR Analysis of Transcripts of NQO1. To dis
tinguishthree types of N001 mRNAin HCT 116 sublines,purifiedmRNA
was reverse transcribed into cDNA, then a target fragment of cDNA from nt
298617 was amplified by PCR. The primer pair chosen was QR-298 and
C to G which is degenerate (threonine 16), and at nt 747 from T to C
that would replace phenylalanine 233 with leucine. The DNA so quence of pDT2O also matched the same portion of the sequence of the human liver NQO1 cDNA (nt 9875)with the exception of two base pairs. One substitution was at nt 747 (F to C), the same substi tution as pDT9. The other substitution of C to T at nt 465 which replaces arginine 139 with tryptophan, a site that was not changed in
QR-617 (Table 1). This region represents a 320-base pair fragment including the entire exons 4 and 5, and part of exons 3 and 6 for the full-length mRNA of NQO1or a 206-base pair fragment for the mRNA of NQO1with the deletion
of exon 4. In both cases, possible contamination and 5. PCR amplification from the amplification of
genomic DNA is eliminated since the amplified region crossed introns 3, 4,
was performed as described above. PCR-produced
DNA fragments were purified by ethanol precipitation, digested by AccIII, and analyzed by a gel containing 1% agarose and 3% NuSieve 3:1 agarose. RESULTS NQO1 Activity and MMC Response of All HCT 116 Sublines. NQO1 activities in soluble extracts of HCF 116 and HC1@ 116-R3OA cells, including the reduction of DCPIP, menadione, and MMC, were compared and reported previously (12). Reductive activity contributed by other flavo-enzymes in cell extracts other than NQOI was ruled out by the use of NADH as electron donor and dicoumarol for inhibition (12). Currently, we estimated NQO1 activities of all HCT 116 sublines by DCPIP reduction using NADH as the electron donor. Soluble cell extracts of all MMC-resistant sublines reduced DCPIP at 5% of the activity found in the parent HCT 116 cell extracts (Table 2). The two revertants, HCT 116-R3OArt1 and HCT 116-R3OArt2, recovered 50 100% of the NQOI activity of parent HCF 116 cells. Their difference in NQO1 activity may reflect their degree of reversion. Microsomal fractions of all the sublines contained similar NADPH:cytochrome P450 reductase activities (data not shown). MMC 50% inhibition of

Table 2 Activity of NQOJ in cell extracts of HCT 116, four MMC-resistant sublines, and two revertants, and their 50% inhibition of cell growth values of MMCaReduction
pDT9. The downstream region of the termination codon of both
clones, including the first polyadenylation site, showed mostly match ing sequences with the human liver NQO1 cDNA (15) with less than
0.5% substitutions.Both strandsof the full insert from both clones
were sequenced, and all substitutions took place at complementary base pairs. Identical results were obtained with multiple sequencing of the same clones.
All other clones, two from HCT 116 and four from HO.' 116R3OA, were shorter than the full-length clones by 114 base pairs. Their sequences matched pDT9 except they missed nt 354467which represented the entire sequence of exon 4 of the human liver N001 gene as reported by Jaiswal (24). One of these clones, pDT23 from HCF 116, was selected for further study. Characterization of NQO1 Expressed by E. coli. N001 cx pressed in E. co/i JM1O9 from both full-length clones, pDT9 and
pDT2O, were successfully purified to homogeneity, as indicated by SDS-PAGE, yielding a single polypeptide subunit with a Mr 30,000 (data not shown). These two purified enzymes were designated as
NQO1wE coil and NQO1139E ccli, respectively. Isoelectric focusing

(,@M)HO' Sublines

of DCPIP (nmol/min/mg) 638.023 29.64

IC501' 0.3 1.6

electrophoresis gave isoelectric pH values of 9.50 and 8.35 for NQO1w'@coil and NQ01139E coil, respectively. Kinetic Properties. The Km of DCPIP for NQO1139'@ coil was about 3-fold greater than that for NQO1wE COIl while their Vm@values were similar (Table 3). When menadione was used as a substrate, the
Km and Vmss of both enzymes were similar. Km and values of
0.05HCT 116 0.22HO' 116-R3OA
NADH and NADPH were obtained by using menadione as electron

0.2HCT'116-R25 116-RiO

1.20.32HO'

29.7 9

acceptor. Menadione rather than DCPIP was selected since the latter
requires high concentrations for saturation that was beyond the accu racy of measurement. Km values of NADH and NADPH were about 5-fold higher for NQO1w'@COilthan those for NQO1139E coIi Vm,@, values of NADH for both enzymes were similar while the Vm,@value of NADPH for NQO1w'@cob was approximately 2-fold higher than that for NQO1139E@@0h. MC reduction at pH 6.2 by these two M enzymes followed pseudo-first order kinetics. Saturation of MMC could not be reached up to 2 mM, the highest soluble concentration.

0.35HO' 116-R40

17.5 2

0.1HO' 116-R3OArtl

116-R3OArt2
a The reduction of DCPIP was

450.038

measured at room temperature

0.3.1 0

in a total of 1.0 ml
containing 50 m@.i Tris-HC1 (pH 7.5), 75 ,LMDCPIP, 0.07% BSA, 0.4 mt.i NADH, and 10100-gil extract by decrease in absorbance at 600 nm The cytotoxicity of MMC to cell
HCT 116 sublines was assessed using the colony formation assay (12). Fifty percent
inhibition of cell growth values of MMC were calculated for each cell line by median effect analysis using a program developed by Chou and Chou (18).
b IC5@,,50% inhibition of cell growth.
QUINONE REDUCTASE AND MITOMYCIN C RESISTANCE coliSubstrateNQO,w'@ Table 3 Kinetic analysis ofNQO, clonedfrom HCT 116 and HCT 116-R3OA cells, expressed by Escherichia

COltNQOI

(nmol/mg/min)Km (nmol/mg/min)DCPIP48.38.2850145159.019.41220 (P.M)Vms,,

coilK,,,

(tiM)Vm.@.
146Menadione21.5 40NADH300.0 49NADPH139.3 70MMC28.9

1.3499 4.6283 7

28.8449

224.6 3

6562.7

Rate (@smol/mg/min)25.2

.11.3272.4220 8.2 3

.212.7 5

a Reduction of DCPIP at various concentrations was measured as described in Table 2 using 0.4 mi.@NADPH as electron donor. Reduction

of menadione

at various

concentrations

was measured in a total of 1.0 ml containing 50 mi.i Tris-HCI (pH 7.8), 80 @M cytochrome c, 0.07% BSA, and 0.4 mr@i NADPH by increase in absorbance at 550 nm. Oxidation of NADH and NADPH was measured using 50 @Menadione and 80 ,.@M m cytochrome c as electron acceptors by following the absorbance at 550 am. Reduction of MMC was measured in a total of 1.0 ml containing 50 nmi imidiazole-HCI (pH 6.2), 0.5 mi@i MMC, and 0.5 m@NADH by following the absorbance at 554 nm. Data represent mean E for at least three S experiments involving triplicate assays.
We previously have shown the loss of NQO1 (quinone reductase) activity in the HO' 116 subline, HO' 116-R3OA, to be the main cause of its resistance to MMC (12). Two full-length NQO1 clones (pDT9 and pDT2O) and an exon 4 deletion variant (jDT23) derived from HO' 116 and HO' 116-R3OA cells were studied to determine the mechanism by which this subline loses 95% of its NQO1 activity. The only difference between the two full-length clones, a substitution of codon 139, replace a positively charged amino acid to a hydrophobic uncharged amino acid (arginine to tryptophan) which probably alters the hydrophobicity and the charge of the NQO1 enzyme. The isoelec tric pH shift from 9.50 to 8.35 of NQO1w'@ colt and NQ01139E coil strongly supports the assumption. The DNA sequences of pDT23 and other exon 4 deletion variants suggested that their protein would lose 38 amino acids (from 102139). e believe that our clones are not the W trivial result of copy mistakes during PCR amplification. These three different cDNAs have been repeatedly produced as we have seen during PCR amplification of the cDNA fragment between nt 298617. A unique restriction site for AccIII is present (nt 462467)in the.@ 2.7Kb PCR amplified nt 298617fragment that is derived from the cDNA of the pDT9 type. Thus, digestion by AccIH of the fragment would yield fragments of 166 and 154 bp (Fig. 4). This restriction site is destroyed in fragments that are derived from cDNA with the substi tution at nt 465 and by the deletion of exon 4. By this analysis, the presence of three different cDNAs for NQO1 (two full length and one deletion variant) in HO' 116 cells was demonstrated by the detection Fig. 1. Northern blot analysis of total cellular RNA from cells of HCT 116, four of four DNA fragments (320, 206, 166, and 154 base pairs) from its MMC-resistant sublines of HCT 116, and two revertants of HCF 116-R3OA.Total RNA was separated by a 1.3% agarose formaldehyde gel under denaturing conditions, blotted, PCR amplified products. These data strongly suggest that at least three prehybridized, and hybridized as described by Sambrook et aL (19). Blotted membranes types of NQO1 mRNAs (two in full length and one deletion variant) were probedwith a 1.3-kilobase cDNA fragmentof a NQO, clone (pDT9) which was are routinely present in the HO' 116 cell population. In addition, each labeledwith (a-32P)dCTPby randompriming.For all samples,30 g.@g RNA were loaded per lane. Lane 1, HCF 116; Lane 2, HCf 116-R3OA;lime 3, HO' 116-RiO; Lane 4, HCF type of mRNA may have two different species (1.2 and 2.7 kilobases) 116-R25; Lane 5, HCF 116-R40; Lmie 6, HCT 116-R3OArtl; Lane 7, HCf 116-R3OArt2. due to alternative polyadenylation. Relative to HO' 116 cells, mRNA

The rate of MMC reduction indicated that NQO1 139'@COIl about had 40% of the activity of NQO1w'@coIi Northern Blot Analysis. Transcription of the NQO1 gene in hu man liver was reported to yield up to three different species (1.2, 1.7, and 2.7 kilobases) due to alternative polyadenylation (15). Two spe cies of NQO1 mRNA (1.2 and 2.7 kilobases) were detected in all HO' 116 sublines (Fig. 1). After stripping the blots and reprobing with a 2-kilobase cDNA fragment of human f3-actin, no differences among all of the cell lines were detected in the signal intensity (data not shown). The steady-state levels of the two species of mRNA by radioactivity estimation showed that the expression of total NQO1 in HO.' 116-R3OA was 90%-80% of the NQO1 that was expressed in the parent cell line (Table 4). Except for HC'F 116-R40 cells, the expres sion of NQO1 in other MMC-resistant sublines and the revertants, HO' 116-R3OArt1 and HO' 116-R3OArt2, was maintained at about 80% of the parent HO' 116 cells. Separation of NQO1 mRNAs of the full-length and the shorter deletion variant was not obtained with the Northern blots. Analysis of Different Transcripts of NQO1. With the analysis of PCR-amplified nt 298617 fragments of cDNA by agarose gel electrophoresis, we detected 2 DNA bands at about 320 and 206 base pairs for both HO' 116 and HO' 116-R3OA (Fig. 2). After AccilI digestion, PCR-amplified nt 298617fragments from HO' 116 cells were resolved into a band at 320 base pairs with decreased intensity, a band at 206 base pairs, and two smaller bands close together at approximately 150170 base pairs. AccIlI did not cut the PCR amplified nt 298617 fragments from HO' 116-R3OA cells, and the intensity of the 320- and 206-base pair bands remained unchanged.
Western Blot Analysis. Rabbit polyclonal anti-NQO1w'@ COlt nti a body reacted equally well with purified NQO1wE COIl and NQO1 139'@coli The enhanced chemiluminescence method detected equal signal intensity by dot blots of serial concentrations of the two enzymes (data not shown). Western blots detected a single protein band with a molecular size of 30,000 in each of the crude cell extracts of HO' 116 sublines (Fig. 3). The four MMC-resistant sublines contained substantially lower amounts of NQOI although five times more protein was used for the blots. MMC-resistant sublines con tamed about 5% of the 30,000 protein of the parent HO' 116 cells (Table 4). Cell extracts of the two revertants, HO' 116-R3OArt1 and HO' 116-R3OArt2, contained the 30,000 protein at levels close to that present in HO' 116 cells.

DISCUSSION

1234567
QUINONEREDUCTASEAND MITOMYCIN RESISTANCE C blot?Total Table 4 Quantitative analysis of Northern blots aad Western

cells expressing

NQOsurvived

better

against

higher

concentra
toHCTI16HCTI162.7-kilobase toNQO, mRNA of NQO, relative.2-kilobaseCell

bandHCT linesband

protein relative

bandMr 0.860.040.05

30,000
tions of MMC, and thus were selected as MMC-resistant sublines. The NQO1 protein detected in HCT 116 cells by Western blot may represent two translational products of the two full-length messengers which could not be distinguished. A smaller protein representing exon
1.01.0HCT 1161.0 0.01HCT Il6-R3OA0.930.12 0.01HCT I 16-RlO0.77 0.01HCT I 16-R250.72 0.01HCT I 16-R400.61 0.06HCT 116-R3OArtl0.89 1l6-R3OArt20.80

a For Northern blots,

4 deletion mRNA could not be detected by our polyclonal anti
.03 0.03 0.02 0.09 0.0.72.030.0.74.030.0.66.020.0.75.030.0.81.060.35 0
of the 1.3-kilobase NQO, cDNA

.05 0

probe associated

radioactivity

NQO3wE coil antibody although exon 4 deletion mRNA was detected in all of the sublines. Whether proteins are translated in cells by the deletion mRNA or the expressed protein is unstable could not be determined at the present time. Poor interaction between the deletion protein with the polyclonal antibody was ruled out because NQO1
with ITLRNA bands in the membrane was first measured by a Betascope 603 blot analyzer. After stripping the initial probe and reprobing with a 2-kilobase cDNA fragment of @3-actin, radioactivity of the (3-actinband was measured and used to normalize the the radioactive counts of mRNA bands (1.2 and 2.7 kilobases) of NQO,. The quantity of both bands for each subline is presented as the percentage of the same band appearing in HCT
116 cells. Data represent mean SE of four separate blots. For Western blots, light
emitted from enhanced chemiluminescence associated with protein bands on the mem brane was captured by the Kodak XAR-5 film. The intensity of each band on the film, representing the concentration of NQO, protein, was determined by scanning with a
Beckman DU7O spectrophotometer. The scanning results obtained for each of the six
sublines was expressed as the percentage of the NQO, protein in HCT 116 cells based on the same amount of crude protein used for the analysis. Data represent mean E of three S separate blots while each blot was exposed to three separate films at different times of exposure, and each exposure was scanned three times.

QUINONE REDUC1@ASE AND MITOMYCIN RESISTANCE C
expressed by pDT23 in E. coli JM1O9 interacted well with the polyclonal anti@NQO,wE coil (data not shown). Our data demonstrated that the 95% decrease of NQO1 activity in MMC-resistant HCT 116 sublines could be ascribed mainly to enzyme concentration and activity, and possibly the type of mRNA. We feel that the concentration of NQO1 in all HCT 116 sublines is controlled by several regulatory mechanisms. One mechanism may be transcriptional regulation that determines
I expression, another may be posttranscriptional modifica
4. Taylor, C. W., Brattain, M. 0., and Yeoman, L. C. Occurrence of cytosolic protein
and phosphoprotein changes in human colon tumor cells with the development of
resistance to mitomycin C. Cancer Res., 45: 44224427, 985. 1 5. Willson, J. K. V., Long, B. H., Marks, M. H., Branain, D. E., Wiley, J. E., and Brattain, M. 0. Mitomycin C resistance in a human colon carcinoma cell line
associated with cell surface protein alterations. Cancer Res., 44: 58805885, 1984. 6. Chakrabarty, S., Danels, Y. J. Long, B. H., Willson, J. K. V., and Brattain, M. G.
Circumvention of deficient activation in mitomycin C-resistant human colonic car
cinoma cells by the mitomycin C analogue BMY25282. Cancer Res., 46: 34563458, 1986.
7. Hoban, P. R., Walton, M. I., Robson, C. N., Godden, J., Stratford, I. J., Workman, P., Harris, A. L., and Hickson, I. D. Decreased NADPH:cytochrome P-450 reductase
activity and impaired drug activation in a mammalian cell line resistant to mitomycin
tions that determine RNA splicing, modulate the stability of mRNA, and the rate of protein synthesis. The disparity between the transcription of NQO1 and the concentration of NQO1 in our MMC-resistant HCT 116 cells relative to HCT 116 cells was a result of different regulation of NQO1 in these cells. It has become increasingly important to establish the role of NQO1 in modulating cellular sensitivity to quinone-containing antitumor drugs (25). Several new agents including 3-hydroxyim ethyl-5-aziridinyl-1-methyl-2-(H-indole-4,7-indione)-propenol (EO9), 5-(aziridin-l-yl)-2,4-dinitrobenzamide (CB-1954), and 3-amino- 1,2,4-benzotriazine-1,4-dioxide (SR 4233) are currently in clinical trials (25). The cytotoxic activities of these agents depends on their reduction by NQO1 and other reductive enzymes. E09 and CB-1954 are activated whereas SR 4233 is detoxified by NQO1
C underaerobicbut not hypoxicconditions. ancerRes.,50: 46924697, C 1990.

8. Marshall, R. S., Paterson, M. C., and Rauth, A. M. Deficient activation by a human cell strain leads to mitomycin resistance under aerobic but not hypoxic conditions. Br.
J. Cancer, 59: 341346,1989. 9. Marshall, R. S., Paterson, M. C., and Rauth, A. M. DT-diaphorase activity and
mitomycin C sensitivity in non-transformed cell strains derived from members of a cancer-prone family. Carcinogenesis. 12: 11751 1991. 180, 10. Dulharty, A. M., and Whitmore, G. F. Chinese hamster ovary cell lines resistant to mitomycin C under aerobic but not hypoxic conditions are deficient in DT-diapho rase. Cancer Res., Si: 18601865, 1991. 11. Siegel, D., Gibson, N. W., Preusch, P. C., and Ross, D. Metabolism of mitomycin C
(MC)by DT-diaphoraseDTD):role in mitomycinC-inducedDNAdamageand (
cytotoxicity in human colon carcinoma cells. Cancer Res., 50: 74837489, 1990. 12. Pan, S., Akman, S. A., Forrest, G. L., Hipsher, C., and Johnson, R. The role of NAD(P)H:quinone reductase in mitomycin C- and portiromycin-resistant HCT I 16 human colon-cancer cells. Cancer Chemother. Pharrnacol., 31: 2331, 1992.
13. Begleiter, A., Robotham, E., Lacey, G., and Leith, M. K. Increased sensitivity of quinone resistant cells to mitomycin C. Cancer Len., 45: 173176,1989.
(25). The dual natureof this enzyme, participatingin the activation

and the detoxification

argument for studies
14. Traver, R. D., Horikoshi, T., Danenberg, K. D., Stadlbauer, T. H. W., Danenberg,
P. V., Ross,D., andGibson,N. W. NAD(P)H:quinone oxidoreductaseeneexpres g
sion in human colon carcinoma cells: characterization of a mutation which modulates DT-diaphorase activity and mitomycin sensitivity. Cancer Res., 52: 797802, 1992.
15. Jaiswal, A. K., McBride, 0. W. Adesnik, M., and Nebert, N. W. Human dioxin inducible cytosolic NAD(P)H:menadione oxidoreductase cDNA sequence and local

of several

to develop

antitumor

drugs,

is a strong

full understanding

of its regulatory

mechanisms. ACKNOWLEDGMENTS We are grateful to Dr. James H. Doroshow for making it possible for S-S. P. to conduct some of the cloning work at the Cloning Laboratory, Division of Biology, Beckman Research Institute and City of Hope National Medical

Center, Duarte, CA and Drs. Nicholas critically reviewing this manuscript. Bachur and Merrill J. Egorin for
ization of gene to chromosome 16. J. Biol. Chem., 263: 1357213578, 1988.
16. Lathe, R. Synthetic oligonucleotide probes deduced from amino acid sequence data,
theoretical and practical considerations. J. Mol. Biol., 183: 112, 1985. 17. Rockwell, S., and Kallman, R. F. Cellular radiosensitivity and tumor radiation response in the EMT6 tumor cell system. Radiat. Res., 53: 281294, 1973.
18. Chou, J., and Chou, T. C. Dose Effect Analysis with Microcomputers. Amsterdam: Elsevier Science Publishers BV, 1985. 19. Sambrook, J., Fritsch, E. F., and Maniatis, T. Molecular Cloning: A Laboratory
Manual, Ed. 2, Vol. 1, pp. 1.251.28. Spring Harbor, NY: Cold Spring Harbor Cold

Laboratory Press, 1989.

REFERENCES
1. Crooke, S. T. Mitomycin C: an overview. In: S. K. Carterand S. T. Creak (eds.),
Mitomycin CCurrent Status and New Developments, pp. 14.New York:
20. Sanger, F., Nicklen, S., and Coulson, A. R. DNA sequencing with chain-terminating inhibitors. Proc. Nati. Aced. Sci. USA, 74: 54635467, 1977. 21. Hattori, M., and Sakaki, Y. Dideoxy sequencing method using denatured plasmid templates. Anal. Biochem., 152: 232238, 1986. 22. Smith, D., Martin, L. F., and Wallin, R. Human DT-diaphorase, a potential cancer protecting enzyme. Its purification from abdominal adipose tissue. Cancer Lett., 42:

103112, 1988.

Academic Press, 1979.
2. Dorr R. T., Liddil, J. D., Trent, J. M., and Dalton, W. S. Mitomycin C resistant L1210
leukemia cells: association with pleiotropic drug resistance. Biochem. Pharmacol.,

36:3115-3120,1987.

23. Laemmli, U. K. Cleavage of strucwral proteins during the assembly of the head of bacteriophage T4. Nature (Lend.), 277: 680685, 970. 1 24. Jaiswal, A. K. Human NAD(P)H:Quinone oxidoreductase (NQO,) gene structure and
induction by dioxin. Biochemistry, 30: 1064710653,1991.
3. Rose, W. C., Huftalen, J. B., Bradner W. T., and Schurig, J. E. Characterization of
p388 leukemia resistant to mitomycin C (MMC). Proc. Am. Assoc. Cancer Res., 27:
25. Riley, R. J., and Workman, P. Commentary: DT-diaphorase and cancer chemother
apy. Biochem. Pharmocol., 43: 16571669,1992.

260, 1986.