DOE

EPA
Department
of
Energy
Lawrence Livermore Laboratory
University of California
Livermore CA 94550
   UCID-18599
United States
Environmental Protection
Agency
Office of Energy, Minerals, and
Industry
Washington DC 20460
FPA 600 7-79-173
August 1979
            Research and Development
            Mutagenicity
            Testing in
            Mammalian Cells

            Multiple  Drug-
            Resistance  Markers
            Interagency
            Energy/Environment
            R&D Program
            Report

-------
                RESEARCH REPORTING SERIES

Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into nine series. These nine broad cate-
gories were established to facilitate further development and application of en-
vironmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The nine series are:

      1.  Environmental Health  Effects Research
      2.  Environmental Protection Technology
      3.  Ecological Research
      4.  Environmental Monitoring
      5.  Socioeconomic Environmental Studies
      6.  Scientific and Technical Assessment Reports (STAR)
      7.  Interagency Energy-Environment Research and Development
      8.  "Special" Reports
      9.  Miscellaneous Reports

This report has  been assigned to the INTERAGENCY ENERGY-ENVIRONMENT
RESEARCH AND DEVELOPMENT series. Reports in this series result from the
effort funded  under the 17-agency Federal Energy/Environment Research and
Development Program. These studies relate to EPA's mission to protect the public
health and welfare from adverse effects of pollutants associated with energy sys-
tems. The goal of the Program is to assure the rapid  development of domestic
energy supplies in an environmentally-compatible manner by providing the nec-
essary environmental  data and control technology. Investigations include analy-
ses of the transport of energy-related pollutants and their health and ecological
effects; assessments  of, and development of, control technologies for energy
systems; and  integrated assessments of a wide range of energy-related environ-
mental issues.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.

-------
                                                  EPA-600/7-79-173
                                                  August 1979
              MUTAGENICITY TESTING IN MAMMALIAN CELLS:

                  Multiple Drug-Resistance Markers
       June H.  Carver,  Gerald M.  Adair,  and Daniel  L. Wandres
                    Biomedical  Sciences  Division
                   Lawrence Livermore Laboratory
                      University of  California
                    Livermore,  California   94550
          Department of Energy Contract No. W-7405-ENG-48
                     EPA-IAG-D5-E681-AN and AO
U.S. D.O.E. PROJECT DIRECTOR          U.S. E.P.A. PROJECT OFFICER
G. Stapleton                          G. Rausa
Office of Health Effects Research     Office of Energy, Minerals and
Asst. Sec. for Environment              Industry
U.S. Dept. of Energy                  U.S. Env. Prot. Agency
Washington, D.C.  20545               Washington, D.C.  20460

-------
                                   FOREWORD
    Short-term toxicologic testing, particularly methods for detecting
genetically active substances, has become a keystone of the Environmental
Protection Agency's efforts to detect and evaluate environmental hazards.
Among the most serious hazards to man's future are agents  that damage the
genome, or the information bank of cells and organisms, by mechanisms that
can increase the future population load of inherited structural defects  and
genetic diseases or that can increase the incidence of cancer in the
current generation.  Tests for environmental mutagens play a dual role by
virtue of their ability to detect mutagens and their presumptive ability to
detect carcinogens, owing to abundant laboratory evidence  that most known
carcinogens respond positively to tests for mutagens.

    The Biomedical Sciences Division of the Lawrence Livermore Laboratory
has for the past five years been heavily engaged in the development and
application of short-term tests for genetic toxicology.  Specific examples
are as follows:

    •    basic studies of classes of mutations in mammalian cells in
         culture and methods for selective recovery of mutant cells;

    •    development of suitable strains of cells for detection of forward
         mutations at multiple loci, and of detailed experimental protocols
         for mutagen detection;

    •    development and application of multiple modes of  testing in vitro
         and in vivo for chromosomal damage and misrepair in the form of
         sister chromatid exchange, together with simultaneous correlation
         with mutation induction;

    •    development and application of highly sensitive tests for injury
         to sperm cells in the male testis and to oocytes in the juvenile
         female ovary;

    •    development of automated cytochemical methods for detection of
         rare events of mutation and malignant transformation of cells in
         living animals, including man;

    •    application of cytogenetic and automated cytopathologic methods to
         workforce populations engaged with energy technology, and

-------
    •    in collaboration with our Environmental Sciences Division,
         application of a battery of the foregoing short-term tests to
         complex effluents from coal gasification and oil shale retorting
         technologies.

    The following report describes the development and partial validation
of one of the major tests in our genetic toxicology test battery.  A strain
of cultured Chinese hamster ovary (CHO) cells has been selected that is
capable of recognizing mutations in four independent genes.  All parameters
that must be controlled for reproducible quantitative assays have been
validated with a set of direct-acting mutagens.  Work now in progress will
extend the assay to chemicals requiring metabolic activation and to
partially fractionated complex effluents.  This assay is a leading
contender among reliable mammalian cell mutagenesis tests that should be
applied when appropriate to confirm the bacterial screening tests, and to
verify that the genetic damage suggested by a screening test also occurs in
the cells of higher organisms.
                                       Frederick T. Hatch, M.D., Ph.D.
                                       Section Leader for Cell Biology and
                                       Mutagenesis
                                       Lawrence Livermore Laboratory
                                     111

-------
                                  DISCLAIMER
    This report has been reviewed by the Lawrence Livermore Laboratory,
University of California, and the U.S. Environmental Protection Agency, and
approved for publication.  Approval does not signify that the contents
necessarily reflect the views and policies of the U.S. Environmental
Protection Agency, nor does mention of trade names or commercial products
constitute endorsement or recommendation for use.
                                     IV

-------
                                   PREFACE

    Short-term mutagenesis assays  are expected  to  play  a major  role  in
identifying potentially  toxic or mutagenic  agents  in the environment.   A
tier concept has evolved  that relates the extent and rigor  of the  screening
protocols  to the expected level or extent of human exposure.  While  the
first tier uses lower  organisms for  rapid,  cost-effective prescreening,
mammalian  cells are desirable at the second level  of testing because
microbial  and mammalian  cells differ in  their uptake and metabolic
modification of many mutagens, and there are marked differences  in the
structure, organization  and  repair of procaryotic  and eucaryotic genomes.
Moreover, mammalian assay systems  evaluate  different portions of the genome
and assay different types of genetic damage.  Assays for forward mutation
at several loci might  be  expected  to respond to most types  of genetic
damage, minimizing false  positive  and negative  tests.   The  latter  point is
of major importance; while many pure compounds  have been tested, the
frequency of false positives and negatives  is virtually unknown  for  complex
organic mixtures or crude effluent samples.

    As an rn vitro test  system, Chinese  hamster ovary (CHO) cells  combine
the favorable growth characteristics, e.g., high plating efficiency  and
rapid growth in either monolayer or  suspension  culture, necessary  for
large-scale cost-effective mutagen screening, with a near-diploid  genome
that facilitates genetic  analysis.  Direct  drug-resistance  selection
procedures, such as those for mutants with  alterations  in genes  coding for
various salvage enzymes  in purine  or pyrimidine metabolism, allow
quantitative mutagenicity testing  at specific well-defined  genetic loci.
We have developed a CHO  cell line  that can  be monitored for forward
mutations simultaneously  at  four gene loci: the autosomal recessive  aprt
(adenine phosphoribosyltransferase)  and _tk  (thymidine kinase) genes,  the
widely used X-linked hgprt (hypoxanthine-guanine phosphoribosyltransferase)
gene, and the co-dominant gene for Na-K-ATPase.  This multiple-marker  assay
can now be used to assess two important  areas:  (1)  determination of  the
quantitative dose-response relationships for toxicity and mutagenicity
observed with compounds  requiring metabolic activation  and  known to  be
present in environmental pollutant mixtures, and (2)  development of  methods
to allow activation and  assay of crude mixtures and effluents after  only
limited fractionation  or  purification.

-------
                                   ABSTRACT

    As a  first  step  in  the  development  of a  multiple-marker,  mammalian-cell
mutagenesis  assay  system, we  have  isolated a Chinese  hamster  ovary (CHO)
cell  line  that  is  heterozygous  for both the  adenine
phosphoribosyltransferase (aprt) and  thymidine  kinase (tk)  loci.
Presumptive  aprt +'~ heterozygotes with intermediate  levels of APRT
activity were selected  from unmutagenized CUO cell populations on the  basis
of resistance to low concentrations of  the adenine analog,  8-azadenine.   A
functional aprt +'~  heterozygote with ~50%  wild-type APRT  activity was
subsequently used  to derive sublines  that were  also heterozygous  for the  tk
locus.  Biochemical  and genetic characterization  of one  such  sub line,
CHO-AT3-2, indicated that it  was indeed heterozygous  at  both  the  aprt  and
^k loci.  Chinese  hamster ovary (CHD) cell lines  heterozygous at  both  the
adenine phosphoribosyltransferase  (aprt)  and thymidine kinase (tk)  loci
were  used for single-step selection of  spontaneous and induced mutants
resistant to 8-azaadenine (AAr), 6-thioguanine  (TGr),  ouabain (OUAR),
or 5-fluorodeoxyuridine (FUdRr).   Mutation data are reported  for  direct
mutagens (EMS, ethyl methanesulfonate;  MNNG,  N-methyl-N'-nitro-N-
nitrosoguanidine;  NQO, 4-nitroquinoline-l-oxide)  and  promutagens  (DMN, di-
methylnitrosamine; BP, benzo(a)pyrene)  activated  by rat  liver homogenates.
Significant dose-dependent  increases  in mutation  frequency were observed
for all four genetic markers  after treatment of CHO-AT3-2 cells with ethyl
methanesulfonate.  Critical plating densities were established for  AAr,
TGr,  and FUdRr; no density  dependence was observed for OUAR.
Expression of mutant phentotypes after  mutation induction with EMS,  DMN,  or
BP was optimal after 2  to 4 d for  AAr,  6  to  8 d for TGr, 3 d  for
OUAR, and 1 to 3 d for FUdRr.  The induced mutant frequencies as  a
function of relative cell survival after  treatment with  EMS,  DMN,  or BP
showed locus-specific differences  in  sensitivity.  Of  59 clonal isolates
resistant to AA and  assayed for APRT  activity,  87% had < 5% wild  type
activity; of 30 TGr  clones  assayed, 83% had  < 5% wild  type HGPRT  activity.
Of 42 FUdRr clones assayed, 98% had < 1%  wild type TK  activity.   Fifty additional
clones selected in medium containing FUdR displayed cross resistance to
5-bromodeoxyuridine  (BUdR)  and trifluorothymidine (TFT)  and all were sen-
sitive to HAT (hypoxanthine-amethopterin-thymidine) medium.   The  jjc  locus
showed the largest mutational response  as a  function  of  cell  survival  after
mutagen treatment.   The rapid expression  kinetics for  FUdRr and the
possibility that the locus  detects a broader  spectrum  of genetic  lesions
than  the other drug-resistance markers  are discussed  in  terms  of  a
sensitive screening  assay for detecting potential mutagens.
                                      VI

-------
    This report was submitted by the Lawrence Livermore Laboratory (Dept.
of Energy, Contract No. W-7405-ENG-48) in partial fulfillment of
Interagency Agreements IAG-D5-E681-AN and AO under the sponsorship of the
U.S. Environmental Protection Agency.  This report covers a period from
January, 1976 to January, 1980; work is still ongoing.
                                     vii

-------
                                   CONTENTS




                                                                       PAGE




Foreword	      ii




Disclaimer	      iv




Preface	       v




Abstract	      vi




Contents	    viii




Lis t of Figures	      ix




List of Tables	      xi




List of Abbreviations	     xii




Acknowledgment	    xiii




    1.  Introduction	       1




    2.  Conclusions and Recommendations	       3




    3.  Materials and Methods	       5




    4.  Experimental Procedures	       9




    5.  Results	      12




    6.  Discussion	      44




References	      52




Glossary	      61
                                    Vlll

-------
                               List of Figures


Number                                                                 Page

  1    Derivation of aprt*'  heterozygous cell lines ..................    12

  2    Derivation of tk+/~ heterozygous cell lines ....................    16

  3    Increase in induced frequency of FUdRr, TGr, AAr, and
       OUAR mutants with increasing concentration (mM) of EMS ...... ...    22

  4    Reconstruction experiments using two aprt*' , hgprt"
       lines with increasing numbers of IB-2 cells ...................    24

  5    The observed frequency of spontaneous mutants selected in
       AG, AA, TG, QUA, FUdR, TFT, or FUdR with increasing density
       of IB-2 cells .................................................    25


  6     Decreasing relative survival (S) of CHO-IB-2 or CHO-AT3-2
        cells with exposure dose (M) of direct mutagens MNNG, 4-NQO,
        and EMS [[[    28

  7     Decreasing relative survival (S) of CHO-AT3-2 cells with
        increasing S-9 or microsome concentrations
        (mg/ml total protein) ........................................    30

  8     Induced AAr mutant frequency as a function of expression
        time before selection with AA ................................    32

  9     Induced TGr mutant frequency as a function of expression
        time before selection with TG ................................    34
  10    Induced FUdRr or OUAR mutant frequency as a function of
        expression time before selection with FUdR or QUA ............   36

  lla   Increase in induced frequency of AAr mutants with decreasing
        relative survival after treatment with DMN, BP, or EMS .......   38

  lib   Increase in induced frequency of TGr mutants with decreasing

-------
lie   Increase in induced frequency of OUAR mutants with decreasing
      relative survival after treatment with DMN, BP, or EMS	 40

lid   Increase in induced frequency of FUdRr mutants after optimal
      expression time with decreasing relative survival after
      treatment with DMN, BP, or EMS	 41

12    Activities of HGPRT, APRT, or TK (relative to wild type) for
      spontaneous and induced mutants	 43

-------
                                List of Tables
Number                                                                 Page

  1   Forward mutation frequencies of parental and heterozygous
      cell lines	  14

  2   Luria-Delbruck fluctuation analysis of azaadenine-resistance
      in IB-2 cells	  15

  3   Luria-Delbruck fluctuation analysis of thymidine analogue
      resistance in AT3-2 cells	  18

  4   Complementation analysis of thymidine analog-resistant strains..  19

  5   Segregation analysis of heterozygosity in CHO-AT3-2	  20

  6   Induced mutant frequency of CHO-IB-2 cells treated with
      direct mutagens	  29
                                    XI

-------
                             List  of  Abbreviations
ct-MEM          alpha-minimum  essential  eagle medium
AA             8-azaadenine
AG             8-azaguanine
AMP            adenosine monophosphate
APRT           adenine  phosphoribosyltransferase,  EC  2.4.2.7
ATP            adenosine triphosphate
BP             benzo(a)pyrene
BUdR           5-bromodeoxyuridine
CdR            deoxycytidine
CHO            Chinese  hamster  ovary
DEAE           diethylamino ethyl cellulose
DFCS           dialyzed fetal calf  serum
DMN            dimethylnitrosamine
EDTA           ethylenediamine  tetraacetic acid
EMS            ethyl methanesulfonate
PCS            fetal calf  serum
FUdR           5-fluorodeoxyuridine
HAT            hypoxanthine-amethopterin-thymidine
HGPRT          hypoxanthine guanine phosphoribosyltransferase,  EC  2.4.2.8
IMP            inosine monophosphate
MNNG           N-methyl-N'nitro-N-nitrosoguanidine
4-NQO          4-nitroquinoline-l-oxide
Na+-
 K+ATPase      Na+- and Reactivated, Mg++-dependent  ATPase,  EC 3.6.1.3
QUA            ouabain  octahydrate
PBS            phosphate-buffered saline
PE             mean plating efficiency
PEG            polyethylene glycol
PRPP           5-phosphoribosyl-l-pyrophosphate
S              mean relative  survival;  surviving cell  fraction
S.D.           standard deviation
SEM            standard error of the mean
TG             6-thioguanine
TFT            trifluorothymidine
TK             thymidine kinase, EC 2.7.1.21
WT             wild type cells
                                     Xll

-------
                               ACKNOWLEDGMENTS:

    The authors thank Drs. W.C. Dewey, T.B. Shows, and R.G. Fenwick for
providing several cell lines used in this study, and Drs. E.W. Branscomb,
L.H. Thompson, M.M. Moore-Brown, and D. Clive for helpful discussions
and/or the provision of unpublished manuscripts relevant to this work.  We
are particularly grateful to Dr. F.T. Hatch for invaluable discussion and
guidance during the course of  this study.  We also thank E.P. Salazar and
M.G. Knize for technical assistance.
                                    Kill

-------
                                   SECTION 1

                                  INTRODUCTION

     In vitro mutagenesis assays are expected to play a major role in
 identifying carcinogenic or mutagenic agents in the environment (8,
 18,65).  Because mechanisms for the generation and repair of
 mutagen-induced damage may vary considerably in procaryotic and eucaryotic
 systems, mammalian mutagenesis assays are needed to confirm and extend the
 microbial data.  The use of multiple genetic markers may increase the
 reliability of such mutagenesis assays for those compounds having diverse
 types of interaction with genetic material.

     The most thoroughly investigated mutant phenotypes of mammalian cells
 are drug-resistant mutants with presumptive alterations in genes coding for
 salvage enzymes in the purine and pyrimidine metabolic pathways (1,5,
 10,13,14,17,22,24,25,26,35,38,45,44,49,52,58,69,70,74,75,78,80,86,92,96)
 Operationally, the definition of a drug resistant phenotype is a mutant's
 ability to survive in the presence of cytotoxic purine or pyrimidine
 analogs.  Analogs frequently used include 6-thioguanine (TG) or
 8-azaguanine (AG), which are metabolized by the salvage enzyme,
 hypoxanthine guanine phosphoribosyltransferase (HGPRT), or 8-azaadenine
 (AA) or 2,6-diaminopurine (DAP), which are metabolized by adenine
 phosphoribosyltransferase (APRT).  Thymidine analogs such as
 5-bromodeoxyuridine (BUdR), 5-fluorodeoxyuridine (FUdR), or
 trifluorothymidine (TFT) selectively kill cells with thymidine kinase (TK)
 activity.  Ouabain (OUA), which specifically inhibits active transport by
 Na+-K+ ATPase, is also used to select drug-resistant mutants (6,19,27).

     The presence of more than one functional allele for autosomal loci in
 mammalian cells makes direct, single-step selection of recessive,
•drug-resistance phenotypes impractical as a means of assaying induced
 mutations.  Mammalian cell mutagenesis studies have, inmost cases, used
 resistance to either 6-thioguanine (TG) or 8-azaguanine (AG) as a
 convenient, well-characterized, and selectable genetic marker to measure
 forward mutation at the X-linked hypoxanthine-guanine phosphoribosyl
 transferase (hgpjrt) locus (1,15,16,23,30,37,38,44-46,50,58,72-76,83,92,96).
 Systematic determination of optimal conditions for the phenotypic
 expression and selection of TGr mutants in Chinese hamster cells
 (15,16,37,48,72,73,74,75,83,92,96) as well as  the development and
 application of a practical, in vitro CHO/HGPRT mutagenesis assay system by
 Hsie and co-workers (29,30,44-47,72-76), has facilitated quantitative
 analysis of mutation induction by a wide range of physical and chemical

-------
agents.  Comparable  selection  of APRT-  or  TK-deficient  mutants requires the
prior derivation of  cells  heterozygous  at  the  autosomal recessive aprt and
tk loci  (17,25,26,52,80,86).   Ouabain (QUA)  resistance  is  a genetic marker
that is  codominantly expressed and  can  be  used to measure  mutation
induction inmost manualian  cells (5,6,19,27,90).   Since the types of
mutational damage, as well as  the mechanisms by which they are produced,
may vary considerably for  different mutagens,  specific  genetic loci may
differ markedly in their mutability by  a particular  agent.   Mutagenesis
assay systems employing multiple genetic markers may be more effective in
detecting a broader  spectrum of mutagenic  and  potentially  carcinogenic
agents.  Extensive validation  and application  of the L5178Y TK+/~ mouse
lymphoma mutagenesis assay developed  by Clive  and co-workers (21-23,25,
26,67),  which permits quantitation  of forward  mutation  at  the  autosomal
thymidine kinase (tk) locus, has revealed  apparent  differences in the
mutagenic potential  of certain agents at the hgprt  and  _tk_  loci.   Several
chemicals that are only weakly or non-mutagenic at  the  hgprt locus appear
to be strongly mutagenic at  the ^k  locus (26).  Operationally, the
relatively short time required for  phenotypic  expression of induced
mutations at the tk  locus  (10,26,67,85) facilitates  its use in mutagenesis
assays.  Since technical disadvantages  limit the utility of the TGr
marker in the L5178Y TK+/~ mouse lymphoma  system (22,23,26,58,67), we
were encouraged to develop a CHO multiple-marker mutagenesis assay system
that could use both  the hgprt  and _tk  loci  and  would  combine and expand the
attributes of the two major  ^ri vitro  mammalian-cell  mutagenesis assay
systems  currently being used.

    In this study, we describe the derivation  of a CHO  cell line  that is
heterozygous for both the  adenine phosphoribosyltransferase (aprt) and
thymidine kinase (tk) loci.  This subline  allows single-step selection of
autosomal recessive AAr or FUdRr mutant phenotypes  as well  as  the more
commonly used OUAR or TGr  genetic markers.  Biochemical and genetic
characterization of  the heterozygous  cell  line, and  dose response data for
mutation induction at these  four genetic loci  by the direct-acting mutagen,
ethyl methanesulfonate (EMS),  are presented.   We also establish optimal
conditions for the phenotypic  expression and selection  of  AAr  and FUdRr
mutants  of CHO cells and present biochemical validation of  the mutant
phenotypes.  Optimal drug  concentrations,  cell plating  densities,  and
expression time requirements are determined for all  four drug-resistance
markers.  Mutation data are  reported  for direct mutagens (EMS, MNNG,  NQO)
and promutagens requiring metabolic activation (DMN, BP).   Finally,  we
discuss  the role and expected  use of  the multiple-marker mutagensis assay
to yield increased sensitivity and reliability in detecting genetic damage
induced  by complex pollutant mixtures of environmental  concern.

-------
                                  SECTION  2

                       CONCLUSIONS AND RECOMMENDATIONS

    The primary objective of  this program  has been  to develop, validate,
and apply multiple-marker, in vitro assays using mammalian  cells  to
quantify the mutagenic effcts of pollutants arising  from energy extraction,
conversion, and use.  The multiple-marker  system, which was initially
validated with data from standardized dose-response  protocols, is  now being
applied to analyzing complex  organic mixtures from crude effluents.
Because most of the suspected mutagens associated with energy by-products
and effluents require metabolic activation, activation techniques  that  can
be coupled with the in vitro mammalian assay are being optimized.
Experiments are being designed to compare  the relative efficacy of organ
homogenates (S9 or microsomes)  versus cell-mediated activation and  also to
test whether our system will yield data at multiple  gene loci which  are
proportional to the relative potency of compounds in other  in vitro  and in
vivo systems.  To further validate the system, paired compounds
(carcinogenic and noncarcinogenic) have been tested.  In this study
(MRC/ICI/NIEHS, Mut. Res. 54: 203, 1978),  results from the multiple marker
assay are being compared with results from a wide assortment of test
systems, including human cells.  The simultaneous use of multiple markers
is expected to test differences in sensitivity or mechanism of damage at
the genetic loci involved, and such a system may increase the generality
and reliability of mutagenesis assays for  compounds having  diverse
interaction with mammalian genetic material.

    Since the mutagenesis assay at mutiple gene loci may be a sensitive
detector of genetic damage induced by complex pollutants of environmental
concern, it should be applied to evaluating potential health effects
associated with effluents from a wide variety of sources.  To determine the
versatility and general applicability of the multiple marker approach to
studies of fractions from crude effluents, chemicals representative of many
different classes should be tested.  Such  compounds, e.g.,  aromatic amines,
alkyl halides, polycyclic aromatic hydrocarbons, nitrogen-containing
heterocyclics, aliphatics, and aromatics,  are frequently found in pollutant
samples and have a broad spectrum of potentially hazardous  interactions
with mammalian DNA.  Many problems are expected in exposing crude effluents
to mammalian cells with an ir» vitro activation system (general toxicity,
inhibitory or synergistic effects among mutagens, chemical  transformation);
in vitro mammalian bioassays must be adapted to accept crude mixtures and
fractionated samples.  These assays for forward mutation at several gene
loci are expected to show a response to a broad variety of  genetic damage,

-------
minimizing false positive and false negative  tests.  The  latter  point  is
important because while many pure compounds have been  tested,  the  frequency
of false positive and negative results  is not presently known  for  complex
organic mixtures or effluent samples.

-------
                                   SECTION  3

                            MATERIALS AND  METHODS

    Cells and culture conditions.  The  parental  cell  line  for  these  studies
was a Chinese hamster ovary (CHO)  cell  line  designated CHO-CIOA,  obtained
from Dr. W.C. Dewey.  One of  the aprt *'   heterozygotes we  derived from
this line, CHO-IB-2, was subsequently used to derive  sublines  that were
also heterozygous for the ^k  locus.  One such subline, designated
CHO-ATS-2, has been further characterized  and used  in a series  of
mutagenesis experiments described  in this  and succeeding papers  (3,16;
Adair & Carver, manuscript in preparation).  Other  Chinese  hamster cell
lines used in this study were the  thymidine  kinase-deficient strains  DON-a3
(obtained from Dr. T.B. Shows), and RJK-92,  originally designated 462-10
(obtained from Dr. R.G. Fenwick).

    Cells were routinely maintained at  37°C  in a-MEM  (K.C.  Biological,
Inc.) supplemented with antibiotics (penicillin/streptomycin)  and 10% fetal
calf serum (K.C. Biological, Inc.).  Cell  numbers were determined with a
Coulter electronic particle counter.  Cultures were periodically  tested and
confirmed to be free of mycoplasma. Under  standard  growth conditions, the
doubling times (T£) of our stock CHO-IB-2  and CHO-AT3-2 cell lines ranged
from 14 to 16 h in monolayer culture and 14  to 18 h in suspension culture.
The doubling time of CHO-IB-2 in suspension  culture containing  dialyzed
fetal calf serum (DECS) was often  20-22 h  with substantial  lag  time after
dilution; CHO-AT3-2 displayed a consistent doubling time under  all culture
conditions, with no significant delays  after dilution.

    Selection of presumptive heterozygotes for the  aprt locus.  Presumptive
aprt *l~ heterozygotes with intermediate levels  of APRT activity  were
selected on the basis of their resistance  to low concentrations of AA.
Unmutagenized CHO-CIOA cells and cells  from  six  clonal sublines of CHO-CIOA
were plated at 10^-10^/100 mm dish, in  a series  of  selection media
containing from 3.0 to 5.0 yg/ml AA.  The  selection media were changed
after one week.  Background growth of cells was  excessive at the  lower AA
concentrations, but discrete colonies were obtained in most of  the cultures
at 4.5 to 5.0 vig/ml AA.  Considerable clonal variation in the  frequency of
these low-level AAr colonies was observed.  A large number  of colonies
were isolated, plated into fresh selection medium containing 5 ug/ml AA to
test whether they were stably resistant, and then secured in liquid
nitrogen.

-------
     Isolation  of HATr  revertants  of TK-deficient mutants.   TK-deficient
mutants of CHO-IB-2 were  selected  in  150  ug/ml  BUdR.   Cells from eight
spontaneous BUdRr  strains  with  barely-detectable levels  of  TK activity
( <1%  that of  wild-type)  were subsequently  plated  into HAT  medium (a-MEM +
FCS  plus  100 yM hypoxanthine, 4pM amethopterin, 50 uM thymidine),  to
select for revertants  which had regained  partial TK activity.   Cells  that
lack TK activity are unable to  utilize  exogenous thymidine  and are  killed
in the presence of amethopterin.   Seven of  the  eight ^k  ~'~ strains
yielded no spontaneous  revertants  in  initial  experiments  (revertant
frequencies <  3 x  10~7).   To obtain HATr  revertants, monolayer cultures
of each of the eight ^k "'  strains were  treated with  EMS (150 pg/ml,  16
hours), subcultured at  2-day intervals  for  8  days  to allow  fixation and
phenotypic expression  of  induced mutations, and then plated into HAT
selection medium (5x10^ cells/100-mm  dish,  12 replicate  dishes/culture).
Five of the eight  lines yielded HATr  colonies,  at  frequencies  ranging
from  3 x 10~7 to  6 x  10~5.

     Preparation of liver  homogenate (S-9  and  microsomes).   Eight week old
outbred male rats  (Simonsen albino, Sprague-Dawley derived)  were injected
i.p. with Aroclor  1254  (200 mg/ml  in  corn oil)  at  500  nig/kg body weight.
Four to five d later,  they were sacrificed  by immobilization in C02
atmosphere followed by  arterial bleeding.   Livers  from > 6  rats  were
dissected aseptically,  washed with cold,  sterile PBS,  placed in PBS
(3 mis PBS/ gm wet liver  weight),  and homogenized  with a Polytron
homogenizer (Brinkmann  Instruments) at  setting  4 for 1 m.   The homogenate
was  centrifuged at 9000 x  g for 15 m  at 4°C and the supernatant  (S-9)  was
quickly frozen and maintained at -80°C  for  < 6  m.  For the  preparation of
microsomes (u), the S-9 supernatant was immediately centrifuged  at
100,000 x g for 1  h.  The  pellet was  gently homogenized, resuspended  in
sterile PBS +  10%  glycerol, and maintained  at -80°C.   Protein
concentration was  determined by the method  of Lowry (63).

     Preparation of cell extracts.  Aliquots of  2xl07 cells  were  taken
from suspension cultures of mutant or wild-type cells  in exponential  growth
phase.  Cells were collected by low-speed centrifugation, washed by
resuspension in cold (0-4°C) phosphate-buffered saline (PBS),  and
recentrifuged.  Each cell  pellet was  resuspended in 1.0 ml  of  cold
extraction buffer  (50 mM KH2P04, 10 mM  2-mercaptoethanol, 1.0% Triton
X-100, adjusted to pH  7),  and the  cell  suspensions were  sonicated on  ice
using a Branson model  S125 ultrasonic cell  disruptor with a microprobe
tip.  Cells were lysed  with four 5-second bursts (power setting  4), with
10-second cooling  intervals.  Resulting cell  lysates were cleared of  debris
by centrifugation  for 30 minutes at 12,000 x  g  (0-4°C), and  the
supernatants were  immediately frozen  and  stored at -80°C.   Protein
concentrations were determined  by  the method  of Lowry  £t al.  (63) using
crystallized bovine serum  albumin  as  a  standard.

     Adenine phosphoribosyltransferase assays.   Adenine
phosphoribosyltransferase  (APRT) activities were assayed using the
radioisotopic/DEAE filter  disc  method of  Fenwick and Caskey  (37), which
measures the rate  of conversion of ^C-adenine  to  ^C-AMP by APRT.
Each 60 wl reaction mixture contained 50 mM Tris-HCl,  pH 7.4,  5 mM  MgCl2,

-------
 2 mM 5-phosphorylribose-l-pyrophosphate (dimagnesium salt), 0.2 mM
 8~  C-adenine (60 Ci/mol),  and 0-48 pg extract protein.  All assays were
 run in duplicate.  Background determinations were obtained by substituting
 extract buffer for cell  extract.   Reactions were initiated by adding cell
 extract to prewarmed reaction mixtures.  Assay mixtures were incubated at
 37°C,  and 20 P 1 samples  were taken after 15 and 30 minutes.  Reactions
 were terminated by spotting the samples directly onto 2.4 cm DE-81 DEAE
 filter discs (Whatman)  that had been pre-treated with 50 U1 of 100 mM EDTA,
 pH 7.0.  After the samples  had dried,  the discs were washed four times wth
 20 ml  volumes of 1:1 methanol-I^O, once with H20, and once with 95%
 ethanol,  to remove unreacted l^C-adenine.  The amount of ^C-AMP bound
 to the discs was subsequently counted  by liquid scintillation.  Reaction
 rates  were constant with time for  at least 30 minutes and proportional to
 the amount of extract protein in  the reaction mixture over the range
 0-48 pg protein.   APRT  activities  of aprt +/~ heterozygotes and AAr
 mutants have been normalized to the mean specific activity of wild-type
 CHQ-CIOA cells,  4.3 nmol/min/mg protein.

     Thymidine kinase assays.   Thymidine kinase (TK)  activities were assayed
 using  the DEAE-disc method  described by Roufa et^ aK  (80),  which measures
 the rate  of phosphorylation of ^C-thymidine by TK.   Each 50 p 1 reaction
 mixture contained 200 PM Tris-HCl, pH  8.0,  2 mM MgCl2,  5 mM ATP, 40 PM
 2-l^C-thymidine  (58 Ci/mol),  and  20-80 pg extract protein.   All assays
 were run  in duplicate.   Background determinations were  obtained by
 substituting extract buffer for cell extract.   Reactions were initiated by
 adding cell  extract to prewarmed  reaction mixtures.   Assay mixtures were
 incubated at 37°C for 10-20 minutes, and reactions were terminated by
 dilution  with 3  ml  of ice-cold 1 M sodium formate,  pH 6.6.   Diluted
 reaction  mixtures were immediately filtered through  2.4 cm DE-81 DEAE
 filter  discs (Whatman),  which were rinsed three times with  cold 1 mM sodium
 formate  and  allowed to dry  before  counting by  liquid  scintillation.
 Background  counts represented less than 0.1% of the  input radioactivity.
 TK activities  of  thymidine  analog-resistant mutants  and ^k  +/-
 heterozygotes  have  been  normalized to  the mean specific activity for
 CHO-IB-2  cells,  171± 15  pmol/min/mg protein.

     Chemicals  and radioisotopes.   All  chemicals were  obtained from Sigma or
 Calbiochem except DMN (Aldrich), 4-NQO (K & K  Laboratories),   BP (Eastman),
 5-phosphorylribose-l-pyrophosphate (P.L.  Biochemicals),  and polyethylene
 glycol  1000  or PEG  (J.T.  Baker Chemicals Co.).    All  were stored at  < 4°C.
 The  mutagens  EMS  and DMN (aqueous),  MNNG (in ethanol),  or NQO (in DMSO,
 final  [DMSO] <0.015%)  were prepared and diluted  with growth  medium
 immediately  before  use.  The  BP stock  (in DMSO)  was stored  < 6 months  at
 room temperature;  the stock was not  filter-sterilized.   Purine analogues
 (AA, AG,  TG)  were dissolved in NaOH and  adjusted to pH  7.4  with HC1;
 ouabain was  dissolved in 100°C 1^0 and immediately diluted  with growth
medium.   The  BUdR,  FUdR,  and  TFT stock solutions (aqueous)  were stored
 frozen  at  -20°C  and protected  from light during all procedures.   All
 solutions  (except BP) were  sterilized  by filtration through a 0.22pm pore
 filter  (Millipore Corp.) before use.   Radioisotopes (8-14C-adenine and
 2-14C-Thymidine)  were obtained from Amersham.

-------
    Statistics.  Where shown, error bars represent one  standard  error  of
the mean (SEM).  To determine the extent of variance from  the mean,  both
the standard deviation of the data points (experimental error) and  the
counting error (sampling or Poisson error) were calculated;  in all  cases,
the larger of these values was used to determine the standard error  (7).
Straight line fits were made by standard linear regression analysis.   The
t test to obtain P values used Scheffe's special d-solution  to the
Behrens-Fisher problem (81); for degrees of freedom, a  conservative  and
minimal value (n-2) was used, with n equal to the number of  data points in
the data set with the smallest number of observations.

-------
                                  SECTION 4
                           EXPERIMENTAL PROCEDURES

    Mutagenesis.  Approximately 18-24 h before  treatment with mutagens,
cells from a recently-thawed culture were plated at  1  to 2 x 10°  per  T25
flask or 3 to 4 x 10° per T75 flask  (Corning).  Treatment with  direct
mutagens was carried out in medium supplemented with PCS (a-MEM PCS)  by
exposing monolayer cultures to fresh medium or  fresh medium containing  the
mutagen ethyl methanesulfonate (EMS), N-methyl-N'-nitro-N-nitrosoguanidine
(MNNG), or 4-nitroquinolinel-oxide (NQO).  After 16-18 h the
mutagen-containing medium was removed, the cells were  rinsed and  dispersed
with trypsin, and 3 to 5 x 10" cells were placed into  suspension  culture
for expression of phenotypes.  Treatment with promutagens requiring
metabolic activation involved a 2 h exposure in a-MEM  (serum-free)
containing S-9 or microsomes plus the following co-factors:  NADPH, 0.37
mM; NADH, 0.93 mM; NADP, 0.87 mM; glucose-6-phosphate, 6.57 mM.   Following
exposure, the cells were washed and allowed a 1 h recovery period  in  a-MEM
PCS.  Approximately 1 to 2 x 10" cells were carried  forward in  monolayer
culture for expression.  After three days incubation,  3 to 5 x  10" cells
were place into suspension culture for continued phenotypic expression  (up
to 10 d) and subsequent replating for mutant selection.

    After mutagen treatment, aliquots of cells were  plated at 200  to  3000
cells/60-mm dish (six replicates per plating) to determine the  surviving
cell fraction (S).  Following 7 d of incubation, surviving colonies were
fixed with methanol, stained with 0.5% crystal violet, and those  colonies
containing 100 or more cells were counted.  The plating efficiencies  of
untreated controls in a-MEM PCS ranged from 0.65 to  0.90; in medium
supplemented with extensively dialyzed PCS (a-MEM DFCS, K.C. Biological,
Inc), plating efficiencies ranged from 0.55 to 0.75.

    Mutant Selection.  To quantitate mutation induction by the  replating
method, mutagenized or untreated (control) cell populations were  incubated
in monolayer and/or suspension culture to allow expression of mutant
phenotypes.  Cell populations treated with direct mutagens were maintained
for 7-10 d in suspension culture; those exposed to promutagens  requiring
metabolic activation were maintained for the first three d in monolayer
culture and subsequently transferred to suspension culture for  4-7 d
additional expression time.  After expression,  the cells were resuspended,
plated, and incubated as follows:

-------
 Mutant          Concentration of      # Cells Plated per   # Days  Incu-
Phenotype     Selecting Drug ()jg/ml)   100-mm petri dish    bation  at  37°
Mr
AGr
TGr
OUAR
FUdRr
BUDRr
TFTr
80
40
4
2200
2
150
3
(0
(0
(0
(3
(0
(0
(0
.59
.26
mM)
mM)
.024 mM)
.00
mM)
.008 mM)
.49
.01
mM)
mM)
3
3
2
8
3
3
3
X
X
X
X
X
X
X
io5
io5
IO5
IO5
IO4
io4
IO4
12-14
16-18
12-14
10-12
12-14
12-12
12-14
Twelve to twenty-four replicate 100-mm dishes containing  16-20 ml  of  drug-
selection medium were plated for each selecting drug.

    With the CHO-AT3-2 cell line, the colony size in extensively dialyzed
serum was enhanced by the addition of deoxycytidine at  10-20 pg/ml  (0.076
mM).  Media containing DFCS was used for mutant selection with AA,  TG,  and
FUdR and included deoxycytidine at 20 ug/ml; FUdR selection also included
uridine at 10 Pg/ml (0.04 mM) to reduce incorporation of FUdR metabolites
into RNA (42,43).

    To quantitate mutation induction by the in situ method, mutagenized
cells were plated in 100-mm petri dishes at approximately 1.2 x 10^
cells/dish and incubated until the number of cells had  increased to the
equivalent of five to six doublings; at that time, the  medium was  carefully
removed and fresh medium containing the appropriate selecting drug  was
added.  The relative increase in the number of cells during the expression
period was monitored by counting (after dispersing with trypsin) replicate
cultures that had been inoculated at the same density as  in the mutation
experiment.  The increase was also determined by plating  100-mm dishes  with
approximately 4 x 10^ cells in 100-mm dishes and counting the number  of
cells per colony-forming group at appropriate intervals (13).  The  maximum
density of cells in the control was about 4 to 8 x 10^  cells per 100-mm
dish (i.e., approximately 0.75 to 1.5 x 10^ cells/cm2 or~10^
colonies).  The actual cell density in mutagenized cultures was somewhat
lower and dependent upon the dose of mutagen.

    To determine the plating efficiency (PE) at the time of mutant
selection, 200 to 400 cells plated' in 60-mm dishes (six replicates  per
condition) were counted.  The observed mutant frequency represents  the  mean
number of mutant colonies per dish divided by the number of viable  cells,
i.e., the number of cells plated per dish, corrected for PE.  The  induced
                                      10

-------
mutant  frequency  is  the  difference between the observed mutant frequency of
treated  cultures  and  the observed  spontaneous  mutant  frequency of controls.

    Luria-Delbruck fluctuation  analysis.   The  pattern and rate of
appearance of drug-resistant  variants  in  cultures  was determined by
fluctuation  analysis  (11,62,64).   Thirty-two independent replicate cultures
were  initiated  from inocula of  50  or  100  cells.  The  cultures  were grown
for 9-10 days (approximately  14-16 doublings)  and  plated into  medium
containing the  appropriate selection  drug.  The  distribution of
drug-resistant  colonies  was analyzed  by the methods of Luria and Delbruck
(64)  and Lea and  Coulson (62)  to obtain estimates  of  the apparent
spontaneous  mutation  rates for  the aprt and ^k loci in CHO-IB-2 or
CHO-AT3-2 cells.

    Cell hybridization.   Cell hybrids  were obtained by PEG-induced fusion,
using the procedure of Davidson and coworkers  (33) as modified by Adair e_t
al. (4).  PEG-treated cultures  were incubated  overnight in a-MEM + PCS,
then  trypsinized  and  replated into HAT medium  to select for  complementing
cell  hybrids.   Cells  that lack  TK  activity are unable to utilize exogenous
thymidine and are killed in the presence  of amethopterin.  Complementation
in cell hybrids between  two thymidine  analog-resistant mutants would allow
colony formation  in HAT  selection  medium,  and  would indicate that the two
mutants did  not share the same  genetic defect.   Cell  hybrids between
HGPRT-deficient TGr cells and each thymidine analog-resistant  strain were
included in  complementation analysis experiments to determine  whether the
TK-deficient mutant phenotypes  were dominant or  recessive, and as a
positive control  for complementation.  Controls  for reversion  for each
mutant strain were also  included,  but  no  revertants were observed in
complementation analysis  experiments.

    Segregation analysis.  Utilizing one  of the  TK-deficient,  BUdRr
strains we had  previously isolated from CHO-IB-2, we  selected  a spontaneous
AAr mutant that lacked both APRT and TK activities.   Cells from this
aprt  '", tk ~/~  strain  were  subsequently hybridized  by PEG-induced
fusion with  a TGr subline of CHO-AT3-2 or TGr  subline of wild-type
(tk +/+) CHO, to  test whether CHO-AT3-2 cells  were indeed  functionally
hemizygous at both the aprt and ^k loci.  Hybrids were selected and grown
to mass culture in HAT medium.  An inoculum of 3x10^  HATr  cells of each
hybrid were  then  grown for approximately  ten generations in  the absence of
further selection and plated into  medium  containing either 80  yg/ml AA or
150 Mg/ml BUdR.   Spontaneous nondisjunction involving the  segregation or
loss  of dominant wild-type alleles  for the aprt  or tk loci from such
hybrids will result in reexpression of recessive AAr  or BUdRr
phenotypes.  The  frequencies of such drug-resistant segregants  should be
much  higher  in  hybrids between mutant  X heterozygotes  than in  hybrids
between mutant  X wild-type cells;  only one wild-type  allele  would have to
be lost for  reexpression  of the mutant phenotype in the former  case,  while
two wild-type alleles would have to be lost in the latter.   Thus,  by
comparing the drug-resistant segregant frequencies in the  above hybrids,  we
could determine whether  CHO-AT3-2  cells were indeed heterozygous  at both
the aprt and tk loci.
                                      11

-------
                                   SECTION  5
                                   RESULTS

Isolation and characterization of  aprt+'~  heterozygotes.   Presumptive
heterozygotes for the autosomal aprt  locus were  selected  from  unmutagenized
populations of CHO-CIOA or clonal  sublines,  on the basis  of  resistance  to
low levels of AA (Fig. 1).  This approach  was originally  used  by  Jones  and
Sargent (52) to isolate aprt heterozygotes from  a different  CHO strain,  but
a subsequent attempt to isolate heterozygotes using  similar  approaches  was
unsuccessful (34).  As reported by Jones and Sargent,  resistance  to AA  is
correlated with a loss of APRT activity and  occurs in  two discrete steps.
The concentration of AA necessary  to  reduce  CHO-CIOA survival  to  10%
(DIQ) was approximately 2.5 iig/ml, compared  to 7-11  Pg/ml reported by
Jones and Sargent.  Using concentrations of  4-5  pg/ml  AA,  which reduced
CHO-CIOA survival to approximately 10~^, we  were able  to  obtain
aprt +/~ heterozygotes with intermediate levels  of APRT activity  (Fig.
1).  Eleven out of twelve presumptive heterozygotes, confirmed to be stably
resistant to 5 ug/ml AA and subsequently assayed for APRT activity, had
specific activities ranging from 47%  to 78%  that of  wild-type  CHO-CIOA.
Three of these presumptive heterozygotes (designated IIA-1,  IB-2, IID-4)
with approximately 50% wild-type APRT activity were  subsequently  employed
in second step selections for spontaneous  and EMS-induced mutants resistant
to higher levels of AA (Fig. 1).   Mutants  resistant  to 20-80 ug/ml AA were
isolated and assayed for APRT activity.  Fifty-five  of the fifty-nine AAr
colonies assayed (93%) had <12% wild-type  APRT activity.   No significant
differences in residual APRT activity were observed  for mutants selected at
20, 40, or 80 ug/ml AA.  On the basis of its favorable growth  and plating
chacteristics, one aprt +'~ strain, CHO-IB-2, was chosen  for further
genetic analysis and use in mutagenesis experiments.

    Forward mutation at the aprt locus.  Spontaneous and  EMS-induced
forward mutation frequencies for the  aprt  locus  for  wild-type  CHO-CIOA-19
and the aprt +'~ heterozygote, CHO-IB-2, are shown in  Table  1.  Observed
frequencies of mutants resistant to 80 yg/ml AA  were at least  several
orders of magnitude higher for the aprt +'~  heterozygote  than  for
wild-type, aprt +'+ cells.  High frequencies of  AAr  mutants were
induced by treating CHO-IB-2 cells with known direct-acting mutagens such
as EMS (Table 1), indicating that  this subline and the AAr marker would
be practical for mutagenesis assays.  Most of the AAr  mutants  assayed had
< 5% wild-type APRT activity.  The spontaneous forward mutation rate at  the
aprt locus in CHO-IB-2 cells was determined  by Luria-Delbruck  fluctuation
analysis to be approximately 3x10"? per cell per generation  (Table 2).
This value agrees closely with the rate reported by  Jones  and  Sargent (52).


                                      12

-------
   Wild-type
   aprt+/+
CHO-CIOA (parental line)
  APRT activity: 1.00
 CIOA-19 (clone of CHO-CIOA)
     APRT activity:  1.00
 (relative to CHO-CIOA:  1.05)
   Presumptive
   aprt+/-
   Heterozygotes
   Presumptive
   aprt~/-
   Homozygotes
        IIA-1
      0.51 ± 0.04
     Spontaneous
      <  0.05(4)

       Induced
   IB-2
0.52 ± 0.02
Spontaneous
 < 0.05(10)
   0.49(1)
  Induced
   IID-4
 0.54 ± 0.02
Spontaneous
 < 0.05(5)

  Induced
                       < 0.05(6)            < 0.05(22)        < 0.05(4)
                         0.10(1)            ~ 0.09(1)         ~ 0.45(1)
                         0.42(1)              0.12(2)
                         0.59(1)

Figure  1.      Derivation of aprt*'  heterozygous  cell  lines, with
               subsequent single-step selection  of  spontaneous  and
               EMS-induced mutants resistant to  AA  (aprt~'~).  Relative
               APRT activities are shown, with the  number of  clones tested
               in parentheses.


    Derivation of heterozygotes for the tk locus.   Although  aprt +/~
heterozygotes  had been readily selected on the  basis of  their  resistance to
intermediate  levels of AA, a similar approach for  isolating  tk_ +'~
heterozygotes  was unsuccessful.  Direct, single-step selection of tk +'~
heterozygotes  in intermediate concentrations of BUdR was impractical due to
excessive background growth of wild-type cells.  Attempts  to isolate
tk +'~  heterozygotes using another thymidine analog, TFT,  yielded
unstable drug-resistant variants at high spontaneous frequencies (2).
Thus, we were  forced to employ an indirect approach similar  to that of
Clive £t £1.  (22),  first isolating rare J* ~'~  mutants of  CHO-IB-2 on the
basis of their resistance to high levels of BUdR,  and  subsequently
selecting HATr revertants of those cells (Fig.  2).  We recovered BUdRr
colonies from  unmutagenized CHO-IB-2 cell populations  in three of five
selection experiments, at an overall frequency  of   2 x 10~'  (Table 1).
All eight spontaneous BUdRr strains assayed, as well as  32/32
mutagen-induced BUdRr or FUdRr isolates, had barely detectable levels
of TK activity (  <1% that of wild-type).  All of these ^k  ~/~  mutants
were stable, HAT-sensitive, and cross-resistant to  BUdR, FUdR, and TFT.
When 8 Jtk ~/~  mutants were subjected to EMS mutagenesis  and  plated into
HAT medium, five  of the eight strains yielded HATr  colonies, at
frequencies ranging from 3 x 10~7 to 6 x 10~5.  To  determine whether
                                       13

-------
                                    Table  1
      Forward mutation frequencies of  parental and heterozygous cell lines
Cell line
Selecting Drug    Spontaneous
                                                          EMS-induced3

CHO-CIOA-19

parental line
+/+ +/+
(aprt > tk )

AA

BUdR

FUdR
_7
3.0 x 10
_7
9.0 + 6.4 x 10 '
-7
5.0 x 10

9.6 +

1.4 +

3.7 +
b
5.2 x 10~6
__b
0.3 x 10~
,b
1.3 x 10~°
IB-2                      AA
aprt heterozygote       BUdR
(aprt +  , tk +^+)      FUdR
                                   3.7 + 0.4 x  10
                                                 -5
                                   2.4 + 0.8 x 10
                                                 -7
                                   7.0 + 7.0 x 10
                                                 -8
                                       6.6 + 0.5 x 10
                                       2.3 + 1.0 x 10
                                       5.0 + 5.0 x 10
                                                     -4
AT3-2                     AA
aprt, tk heterozygote   BUdR
(aprt +/~, tk +/~)      FUdR
                2.7 + 0.2 x 10
                                                 -5
                2.8 + 0.2 x 10
                              -3
                1.6 + 0.2 x 10
                              -3
                                                          1.5 + 0.1 x 10
                                                     -3
                                                          6.1 + 0.3 x 10
                                                                        -3
 Expression time 7-8d.
b
 100-125 pg/ml, 16-18 hr exposure, S'
C160 ug/ml, 16 hr exposure, S ~0.35.
 420 ug/ml, 16 hr exposure, S ~0.30.
                    0.65 - 0.70.
                                      14

-------
                                    Table 2
                    Luria-Delbruck  fluctuation  analysis  of
                      azaadenine-resistance in IB-2 cells
*J
Replicate cultures
Initial number of cells per replicate
Final number of cells per replicate
32
100
1.75 x 107
1
100
1.75 x


107
Cells tested per sampleb; (# samples)    3.6 x  106  (1)         3 x  105  (48)
Number of mutants per sample:
Range
Median
Mean
Variance
Ratio: Variance/Mean
Null fraction
Mutation rate calculated by:
d
Mean method
g
Median method
p
Null fraction method
Q
Maximum likelihood method

0-89
1
3.7
244
65.5
0.47

_7
3.6 x 10 '
_7
3.2 x 10
_7
2.7 + 0.6 x 10
-7
2.4 _+ 0.5 x 10

0-2
0
0.31C
0.38C
1.23°
0.85









Replicate cultures initiated by innoculating Corning T-25 flasks with 100
 cells were subsequently transfered to T-75 flasks.
 Cells were plated at 3 x 10 /100-mm dish in medium containing 80 pg/ml
 AA and incubated for 12 d to allow colony formation.
CThese numbers may be underestimated due to limitations of sample size
 which resulted in many plates with no mutant colonies.
^ef. 11.
eRef. 62.
                                      15

-------
these colonies  actually represented ^k +/~ revertants, a total of  33
HATr colonies were  isolated,  grown to mass culture, and assayed for TK
activity.  TK specific  activities ranged from 7-184 pmol/min/mg protein
(compared  to 171  +_  15 pmol/min/mg protein for wild-type CHO-IB-2), with
most of the revertants  having 10-30% wild-type TK activity.  Further
biochemical characterization  of several of these presumptive tk +'~
heterozygotes (Adair, manuscript in preparation) revealed differences in  rn
vitro thermolability and substrate concentration optima for TK, suggesting
that they  indeed  represent  structural gene mutations.  One of these
tk +/~ heterozygotes, CHO-AT3-2, was chosen for further genetic analysis
and use in mutagenesis  assays.
               Wild-type
               tk+/+
          CHO-IB-2
     TK activity: 1.00 ±0.09
               Presumptive
               tk-'-
               Homozygotes

               Presumptive
               tk+/-
               Heterozygotes

               Presumptive
               tk-/~
               Homozygotes
 SBUdR'-2
  <0.01

     I
  AT2-4
0.49 ± 0.06
     I
Spontaneous
< 0.01  (18)
 SBUdRr-3
 <0.01
  AT3-2
0.36 ± 0.04
    I
Spontaneous
< 0.01 (23)
  0.15(1)
Figure 2.     Derivation  of  tk+'~  heterozygous cell lines by isolating
              tk~/~ mutants  of  CHO-IB-2,  with subsequent EMS mutagenesis
              and HAT  selection of tk+'~  revertants.  Spontaneous mutants
              resistant to FUdR were recovered in a single step; relative
              TK activities  are shown,  with the number of clones tested in
              parentheses.
                                      16

-------
    Forward mutation  at  the  tk  locus.   Although  BUdRr  or  FUdRr, ^
mutants were rarely recovered from uninutagenized CHO-IB-2 populations
(Table 1), such variants were readily  obtained from  CHO-AT3-2  at  very  high
spontaneous frequencies  (2-3 x  10"-*).   To determine  whether  expression of
^k ~/~, thymidine analog-resistant phenotypes in CHO-AT3-2 indeed
represented spontaneous, randomly occurring "mutational"  events,  and
whether the high frequencies of  these  variants in unmutagenized populations
reflected a high spontaneous rate of mutation at the tjk locus, we performed
a Luria-Delbruck fluctuation analysis  experiment.  Cells  from  each
replicate culture were plated into media containing  BUdR,  FUdR or TFT,
allowing us to determine the spontaneous rate and pattern of occurrence of
drug-resistance for each selection agent.  In each case,  as  shown by data
in Table 3, the variance for replicate, independent  cultures was
significantly greater than that  for multiple platings  of  a single culture.
These results indicate that Jtk ~/~, thymidine analog-resistant variants
arose spontaneously and  randomly in CHO-AT3-2 populations, suggesting  that
acquisition of the thymidine analog-resistant phenotype represented a
"mutational" event occurring prior to  drug selection and  was not  merely an
adaptive response to  the selecting drug.  Calculation  of  apparent forward
mutation rates by several methods (Table 3) indicated  that ^k  ~/~,
thymidine analog-resistant variants arise at a rate  of 2-3 x 10~^ per
cell per generation.  These estimated  mutation rates are  approximately
three orders of magnitude higher than  those for  single-step, forward
mutation at the hgprt or aprt loci in  the same cell  line  (Table 2 and
unpublished results), and are simlar for BUdR, FUdR, and  TFT selections.

    Complementation analysis of  thymidine analog-resistant strains.  The
thymidine analog-resistant strains we  have isolated  characteristically lack
appreciable TK activity  and thus are killed in HAT selection medium.   To
determine whether the HAT-sensitive phenotypes of these mutants were
dominant or recessive, and to determine whether  all  shared the same genetic
deficiency, a series of mutants were subjected to complementation analysis
by fusion in pairwise crosses.  Complementation  analysis  of  BUdRr,
FUdRr, and TFTr derivatives of CHO-AT3-2 with other  TK-deficient  mutant
strains (Table 4), indicated that all  of these mutants behave  recessively
and fail to complement each other or other tk ~'~ Chinese  hamster cell
lines such as DON-a3 or RJK-92 (V-79 462-10).  Thus, all  of  these mutants
belong to a single complementation class and presumably reflect mutational
lesions at the same genetic locus.  The DON-a3 cell  line  has been
previously used to map this presumptive, structural  gene  locus for TK  in
both the mouse (60) and human (36) karyotypes.

    Segregation analysis of heterozygosity in CHO-AT3-2.   Further evidence
that CHO-AT3-2 cells were indeed heterozygous at both  the  aprt and Jrk  loci
was obtained by comparing the frequencies of AAr or  BUdRr  segregants
from mutant X CHO-AT3-2  and mutant X wild-type cell  hybrids.  For both
loci, the frequencies of drug-resistant segregants from mutant X  CHO-AT3-2
hybrids were at least 2-3 orders of magnitude greater  than those  from
mutant X wild-type cells (Table  5), as would be  expected  in  the case of a
double heterozygote with a single wild-type allele at  both the aprt and ^k
loci.
                                      17

-------
                                                                Table 3
                                           Luria-Delbruck  fluctuation  analysis  of  tbymidine
                                                   analogue resistance  in AT3-2  cells
oo
Replicate cultures
Initial cell number per replicate
Final cell number per replicate
Cells tested per sample (# samples)
Number of mutants per sample:

Range
Median
Mean
Variance
Ratio: Variance/Mean
Mutation rate calculated by:
Mean method3
Median method
c
Accumulation with time , exp A

exp B
31 1
50 50
2.7 x 107 2.7 x 107
9.0 x 104(1) 9.0 x 104(12)

138-293
180
188.6
1597
8.47

2.1 +
3.5 +.
2.7 x

1.7 x
BUdRr
144-210
174
173.1
254
1.47

0.5 x 10-4/cell/gen
0.4 x 10~4/cell/gen
10~4/cell/gen
—L
10 /eel I/ gen

83-225
124
134.6
1266
9.41

1.6 1
2.6 +
1.8 x

1.2 x
FUdRr
111-150
120
40.9
321
2.61

0.4 x 10~4/cell/gen
0.3 x 10~4/cell/gen
10~4/cell/gen
— L
10 Vcell/gen
TFTr
77-205
127
122.7
852
6.53

1.5 +
2.6 +
1.8 x

1.2 x

99-141
117
120.0
206
1.71

0.3 x 10~4/cell/gen
0.3 x 10~4/cell/gen
10~4/cell/gen
—L
10 /cell/gen
          64.
    bRef. 62.
    °Ref. 82.  Experiment A and B separated by three additional days of expression time.

-------
DON-a3
RJK-92
SBUdRr-l
SBUdRr-2
SBUdRr-3
ETFTr-2
BUdRr AT3-2
FUdRr AT3-2
TFTr AT3-2
STGr-l
                                                     Table  4
                         Complementation analysis of thymidine  analog-resistant  strains8

                                                                                   BUdSr   FUdRr   TFl'r
                DON-a3b    RJK-92C    SBUdRr-l   SBUdRr-2   SBUdRr-3    ETFTr-2     AT3-2   AT3-2   AT3-2 STGr-ld
 Complementation (+) or noncomplementation (-) based upon growth of hybrids  in HAT  selection medium.
 TK-deficient strain of DON cells, isolated by Westerveld et^ al_. (98).
CTK-deficient strain of V-79 cells (originally designated 462-10), characterized  by Roufa £t al_.  (80).
 HGPRT-deficient strain of CHO isolated on the basis of its resistance  to 4pg/ml TG.

-------
                                      Table 5
                Segregation analysis of heterozygosity in CHO-AT3-2
                                         Frequency of drug-resistant
                                                               o
                                             segregant colonies
             Hybrid
     AA1
             BUdRr
AAr, BUdRr Mutant
(aprt'/', tk~'~)
Expt. 1
Expt. 2
X Wild-type CHO
+/+ +/+
(aprt , tk )
4.8 + 1.5 x 10~6 4.3 + 4.3 x 10~7
7.2 + 1.7 x 10~6 13.4 + 7.7 x 10~7
AAr, BUdRr Mutant X CHO-AT3-2
(aprt'7", tk"7")  (aprt+/~, t
       Expt. 1
       Expt. 2
1.4 + 0.1 x 10
0.7 + 0.1 x 10
-3
-3
2.5 + 0.5 x 10
5.0 + 0.7 x 10
-4
-4
 Spontaneous frequencies determined after~10 population doublings, by plating
 hybrid cells into media containing the appropriate selecting drug.  Segregant
 frequencies have been corrected for P.E. (0.50-0.65).
                                      20

-------
    Assay of specific locus mutation using multiple markers.  Using
CHO-AT3-2 cells, it is possible  to quantitate mutation induction at  the
autosomal aprt and ^k loci, as well as in the genes involved  in
QUA-resistance or at the X-linked hgprt locus.  We have defined optimal
conditions for the selection of  TGr, AAr, OUAR, and BUdRr or
FUdRr mutant phenotypes.  Establishment of selection parameters such as
maximal cell plating densities and expression kinetics for each drug
selection marker are described (16).  Maximal TGr mutant frequencies are
attained by 6-8 days of expression and plateau at that level.  Full
expression of AAr or OUAR mutants requires only 2-4 days, with no
apparent loss of mutant cells by longer expression times.  Expression
kinetics for resistance to FUdR  are more complex with mutagen-specific
differences in optimal expression time (16; Adair & Carver, manuscript in
preparation).  Dose response data for mutation induction after 8 d
expression by the direct-acting  mutagen, EMS, in the genes for APRT, TK,
HGPRT, and Na+/K+ ATPase (OUAR)  in CHO-AT3-2 cells are presented in
Fig. 3.  Significant dose-dependent increases in mutation frequency were
observed for all four drug-resistance markers; in each case there is a
linear dose-response relationship between mutation frequency  and EMS
concentration.  The slopes for mutation induction as a function of applied
molar dose were: FUdRr, 1.31 + 0.04 x 10~3/mM; AAr, 4.14 + 0.05 x
10~4/mM; TGr, 5.21 +0.14 x 10~4/mM; OUAR, 1.27 + 0.09 x 10-4/mM.
These data indicate that the gene for TK may be much more mutable than the
genes for APRT, HGPRT, or Na+/K+ ATPase (OUAR).
                                       21

-------
         600
         400 —
      I
         300 —
       a>

      I
      •*

      I
        200 —
        100 —
                                      [EMSl.mM
Figure 3.  Increase  in induced frequency of FUdRr (-*-), TGr(—*	)AAr
(--•—), an(j QUA^  ( v  )  mutants  with increasing concentration (mM) of EMS.
Different experiments  are  represented by open and closed symbols.  The  linear
regression of  the straight line fit (Y = bg + b^ X, where X=  EMS  , mM)
has the following parameters  (+ 1  S.D.):
    FUdR

    TGr:
    AA1
    QUA
       R
               0
-0.04 + 0.11 x 10
-2.96 +_ 2.47 x 10

 0.02 + 1.37 x 10
bfl = -1.65 + 2.38 x  10
-3

-5
  «
  i
-5.
  i
-5
;  b  = 1.31 + 0.04 x 10

;  b  = 5.21 + 0.14 x 10
     = 4.14 + 0.05 x 10
   ,  = 1.27 + 0.09 x 10
                                                -3

                                                -4
                                                -4
                                                -4
                                       22

-------
    Cell Density Analysis.  The number of wild  type  cells  plated  affects
the survival of drug-resistant mutant cells  in  the selecting  drug.   As
previously reported  (13), reconstruction experiments  with  wild  type  CHO
cells showed that a  critical density for AGr mutant  recovery  in the
selecting drug could be as low as 3 to 4 x 104  cells/cm2.   For  CHO-IB-2,
a similar reconstruction experiment (Fig. 4) indicated  that the mean
critical density (MCD5Q, or density of cells yielding 50%  mutant
survival) for recovery of an aprt ~'~, hgprt" mutant  was approximately
1.3 x 104 cells/cm2  in 80 ug/ml AA and 3.0 x 104  cells/cm2 in
40 ug/ml AG.  The recovery of two such spontaneous aprt '~, hgprt~
mutants was reduced  by 50% at approximately  7.5 x 10-* cells/cm* in
4 vig/ml TG.

    Reconstruction experiments such as these measure  the response of a
single mutant phenotype, i.e., one prototype mutant  clone  isolated from a
selection experiment and grown to a population  sufficiently large for
testing.  However, a number of independent mutational events, and thus a
variety of mutant phenotypes, contribute to  the frequency  of  spontaneous
mutants.  In Fig. 5, an experiment designed  to  test  the optimal plating
density for maximum  recovery of spontaneous mutants  is  shown.   The observed
spontaneous mutant frequencies of AAr, AGr, TGr,  OUAR and
FUdRr/BUdRr/TFTr are compared as a function of  the number  of
CHD-AT3-2 or CHO-IB-2 cells plated.  The cell density ranged  from 1  x 105
to 1 x 106 cells per plate (1 x 104 to 1 x 105  cells  per plate  for
pyrimidine analogs), corresponding to approximately  1.9 x  10-* to  1.9 x
104 cells/cm2 and 1.9 x 102 to 1.9 x 10^ cells/cm2, respectively.

    In Fig. 5a, recovery of AAr mutants decreased with  increasing cell
density (50% decrease at about 6.5 x 10^ cells/plate) corresponding  to
the metabolic cooperation observed in Fig. 4.  The experiments  described
(Table 6, Figs. 8, lla) used 80 pg/ml AA and 3 x  105  cells/plate, a  drug
concentration and cell density that allows maximal recovery of  spontaneous
mutants.  As shown in Fig. 5a, selection with either  30 or  40 gg/ml  AG was
not affected by cell concentrations  up to 1 x  10° cells/plate
(1.85 x 104 cells/cm2).  The dose response data described  in  Table 6
were obtained with 40 pg/ml AG and 3 x 105 CHO-IB-2  cells/plate.

    For both CHO-AT3-2 and CHO-IB-2, the optimal  cell density for recovery
of TGr mutants (Fig. 5b) is approximately 1 to  2  x 10^  cells/plate;
mutant recovery was  reduced by 50% at 8 to 9 x  10-* cells/cm2.   No
significant increase in recovery was seen at either  5 x 104 or  7.5 x 104
cells/plate (data not shown).  The data in Table  6 and Figs.  9  and lib
correspond to selection with 4 yg/ml TG and 2 x 10^ cells/plate.  For
OUAR mutants, no density effect was observed up to 8  x  105  CHO-IB-2
cells/plate (Fig. 5b).  This was tested not only  with new  spontaneous
mutants but also in  a reconstruction experiment with  200 OUAR
cells/plate; the CHO-IB-2 cell density ranged from 1  x  105  to 4 x 106
cells per plate.  The resulting plating efficiency (data not  shown)  showed
no significant change as a function of the number of  CHO-IB-2 cells,
although smaller colonies were observed when 2  and 4  x  106  cells were
plated.  The frequencies of OUAR mutants shown  in Table 6,  Figs.  lOc and
lie correspond to selection with 3mM OUA and 8 x  105  cells/plate.


                                     23

-------
                                       106
                                Wild type cells/plate
10'
Fig. 4  Reconstruction  experiments  using two aprt~/~,  hgprt" sublines
derived from CHO-IB-2 cells  (one  line  selected first in AA and then in TG,
open symbols, and one line selected in AA followed by  AG, closed symbols).
The resistant population was  plated in 100-mm petri dishes at different
IB-2 cell densities  (O= 40 pg/ml AG,  D= 80 yg/ml AA, and A= 4 ug/ml TG).
                                      24

-------
              10
                -4
            .
           O>
           0.
           u
           0>
              10
                -5
                                                         t
                                                        a .
                                               AA80
           AA40
           £  10~4
           c
           CD
           .Q
           O
              10
                -5
   _:£__?_  5   AG30
$-!-§--- -5— -5-
                                                AG40
                 I    I    I    I
                      1   23456789  10

                            Wild type cells/plate, X 105
                                                                5a
Fig.  5   The observed frequency of spontaneous mutants selected from
increasing density of CHO-IB-2 cells (closed symbols) or CHO-AT3-2 cells
(open symbols).   (a) 30 ug/ml AG (—•--) or 40 pg/ml AG (—O—); 40 pg/ml
AA ( -•—); or 80 wg/ml AA (—P— ).  (b) 4 wg/ml TG (-A-^ , —A— )
or 3  mM QUA (-»-).  (c) 150 wg/ml BUdR (-o—); 3 y g/ml TFT (—O—);
2 ug/ml FUdR (—A—).  Error bars are 1 SEM.
                                  25

-------
8
_£
.O
u

Q>
3  ««-5
ff
(D
*J
i
"S
.O
O
    10-
    10
      -6
                    I     I     I     I
123456789

         Wild type calls/plate, X 105
                                                             10
                                                                       5b
  8
  8
 .O
 O
•s -_ •»
d mutant frequency per vk
* —
3 C
J.
I I I I I I
" Q 5 Q - - Pi
y :: Q y Q
- S ^ T 9 ? 0
-1- T A A X .1
- £ * ~ 4
1 1 1 1 1 1

1 1 1
BUdR 150
TFT 3

FUdR2
I ! I

I-
c .
4
i
      234567

         Wild type cells/plate, X 104
                                                             10
                                                                      5c
                               26

-------
    The high  spontaneous  frequency  of  tk'~ mutants  allows  platings  below
the range where metabolic cooperation  is  thought  to  occur.   In  Fig.  5c  the
frequency was constant  over  the range  of  1 x  10^  to  1  x  10^  cells/plate
(2 x  102 to 2 x 103 cells/cm2); data in Figs.  lOa, lOb,  and  lid are
the result of 3 x  104 cells/plate in 2 ug/ml  FUdR.

    Multiple Marker Dose Response for  Direct  Mutagens.   The  relative

survival (S), or the fraction of surviving CHO-IB-2  and  CHO-AT3-2  cells as
a function of applied mutagen dose  of  MNNG, NQO,  or  EMS,  is  given  in Fig. 6.
Table 6 shows the  induction  of mutants at three genetic  markers (AAr,
AGr/TGr, and OUAR).  Detailed dose  response data  on  EMS  and  MNNG were
previously reported for CHO  cells (29,30,45);  our emphasis was  primarily to
establish empirically how data for  these widely-used,  direct-acting  mutagens
varied as a function of cell line and  different culture  conditions during
expression.  Results in Table 6 are pooled from experiments  carried  out
over 1 y using a single lot  of FCS; data for  experiments  using  a different
lot of serum were  consistently elevated and are not  shown.

    When shown as  a function of relative cell  survival for all  three
markers, the response in Table 6 was not always linear at lower cell
survival (<0.50).  The nonlinearity might reflect unequal growth  rates of
mutant and wild type cells after higher mutagen doses, resulting in  an
apparent decline in mutant frequencies.  Obtaining mutation  data by  the
replating protocol assumes that both mutant and wild type cells grow at
approximately the  same rate  during expression  in  suspension  culture.  We
tested this assumpution by performing  parallel experiments employing either
(a) expression of  cells in suspension  and replating, or  (b)  in  situ
expression and selection of  cells plated at a  low cell density  (to avoid
metabolic cooperation).  Results from  the in  situ experiment are also shown
in Table 6.  Differences in  the AAr mutant frequencies in the replating
and iji situ protocols after  5 to 6 doublings  (in  situ) or 5  to  7 doublings
(replating) were significant only for  mutants  induced  by  EMS.   However,
both the TGr and AGr mutant  frequencies for in situ  selection were
significantly higher for both EMS- and MNNG-induced mutants.
Colony-splitting occurring at the time drug medium was added is a
possibility, but the correlation of MNNG- and  4-NQO-induced  AAr mutants
selected either rn situ or replated indicates  that the magnitude of
colony-splitting was variable and small in magnitude,  if  it  occurred at
all.  Therefore, at least for CHO-IB-2 mutants induced by EMS and MNNG,  a
subpopulation of HGPRT" mutants may be at a growth disadvantage in
suspension.  The growth disadvantage could account for the nonlinearity
observed at low survival, where extended growth delays resulting from
toxicity obscure the fate of the surviving mutant subpopulation.  The
delays may negate  the assumption that  mutant  and wild  type cells have
similar growth properties over the expression  period.
                                      27

-------
      1.00p
        10
                                   Exposure dose, M
Fig. 6   Decreasing  relative survival  (S) of CHO-IB-2  (•,*,•)  or  CHO-AT3-2
(T, *) cells with exposure dose  (M)  of  direct mutagens  MNNG,  4-NQO,  and
EMS.  Solid and open symbols represent  separate  experiments.   CHO-IB-2
cells were treated for  18 h, CHO-AT3-2  cells for 16 h.
    Expression may also be carried out in monolayer  culture; with  expres-
sion in monolayer culture, linearity of  the dose  response  has  been re-
reported under conditions of low survival for some mutanges  (29,30) but  not
others (61,62).  The multiple marker mutagenesis  assay  reported here
requires a large number of cells per dose point,  e.g. about  5  x 10? for
spontaneous and low dose determinations, if optimal  statistical param-
eters are desired.  The need to maintain cell density below  a  critical
level in monolayer culture during a prolonged expression period (1,13)
makes monolayer expression impractical for these  large-scale experiments.
Transfer of the cells  to suspension culture for two  or more  doublings
before drug selection  appears to be sufficient to minimize cell-to-cell
interaction before selection (Carver, unpublished data).

    Metabolic Activation.  At a dose of  9 mM DMN, maximal  induction of
mutation for all four markers was obtained with concentrations < 1.6 mg/ml
S-9 protein; a similar increase in DMN-induced mutants  as  a  function of
cell survival was observed when microsomal activation (<0.75 mg/ml) was
used.  When S-9 protein concentrations were > 2.0 mg/ml S-9  or  > 1.0 mg/ml
microsomes, toxicity was not increased (Fig. 7) and  the induced mutant
frequency (data not shown) remained level or decreased.
                                      28

-------
                 TABLE  6  Induced  mutant  frequency  of  CHO-IB-2  cells




                            treated with direct mutagens
pg/ml
EMSb




(In Situ)
N00b





(In Situ)
(In Situ)
MNNGb




(In Situ)

100
125
175
225
275
125
0.30
0.40
0.50
0.55
0.65
0.70
0.65
0.70
0.07
0.08
0.09
0.11
0.14
0.11
S
0.74
0.30
0.17
0.04
0.02
0.30
0.96
0.83
0.77
0.66
0.49
0.15
0.49
0.15
0.75
0.44
0.26
0.12
0.05
0.12
AAr x 10
2.4
7.3
9.7
7.4
11.3
20.0
0.1
0.6
2.2
3.5
5.4
6.2
3.0
5.6
1.1
2.2
2.2
2.7
2.8
5.9
1 °
± °
± °
± °
± °
+ 3
± °
± °
± °
± °
± °
± °
+ 1
+ 2
1 °
± °
± °
± °
± °
+ 2

-4
.2
.3
.5
.4
.6
.8
.01
.04
.2
.4
.4
.4
.2
.2
.1
.1
.2
.2
.2
.4
AGr x
4.7
14.1
20.5


108.8
0.2
1.5
2.6
10.1
16.8
19.0
18.8
35.8
0.8
6.7
7.1
7.1
7.7
49.5
-4 r -4
10 TG x 10
+ 0.2
+ 0.8
± 1-2
-
-
+ 16.3
+ 0.0
+ 0.1
+ 0.2
+ 0.8
± 1-1
+ 2.0
+ 5.6
+ 10.7
+ 0.1
+ 0.4
+ 0.5
+ 0.5
+ 0.7
+ 14.9
1.1
5.0
7.0
7.7
8.4
34.8
0.4
0.7
1.7
3.1
4.9
6.3
9.3
10.4

0.6
0.8
1.0
1.1
6.8
+ 0.1
± 2-5
+ 0.4
+ 0.3
+ 0.4
+ 2.6
+ 0.31
+ 0.5
+ 0.2
+ 0.3
+ 0.5
+ 0.4
+ 1.2
+ 2.1

+ 0.1
+ 0.1
+ 0.1
i o.i
•f 2.6
OUAR
2.5
10.0
9.6
8.9
13.1

0.5
1.5

3.0
3.9
5.4


0.3

1.6
2.1
2.4

x 10~5
+ 0.2
+ 0.5
+ 0.5
+ 0.5
+ 0.6
-
+ 0.04
+ 0.1
-
+ 0.3
+ 0.4
+ 0.5
_
_
+ 0.05
-
+ 0.1
+ 0.2
+ 0.2
"•
aSpontaneous frequencies were as follows:   M  1.2 x 10  ;  AG  1.0 x 10  ;




TGr 4.6 x 10~5j OUAR 5.1 x 10~6.




^Treatment time was 18 h.
                                  29

-------
  1.00
5
E
3

(£
  0.10
                                                                      2.00 [S-9]
                                                                      1.00  l/i]
 Fig.  7   Decreasing  relative survival (S)  of CHO-AT3-2 cells with increasing
 S-9  (open  symbols)  or  microsome (y,  solid symbols) concentrations (.mg/ml
 total protein).   The cells  were treated for 2 h with the promutagens DMN at
 9 mM (o,«)  and BP at 0.04 mM (A,A).


      When activation of BP by S-9 at  concentrations of 0.50 to 2.0 mg/ml was
 compared to activation by microsomee at 0.25 to 1.0 mg/ml, the data
 indicated  that BP was  most  effectively metabolized to mutagenic products by
 microsomes  at  <  0.50 mg/ml; microsome concentrations> 0.50 mg/ml
 consistently  lowered the optimal yield of BP-induced mutants.  Although we
 observed only minimal  induced mutation from BP activated by S-9 at
 > 0.5 mg/ml,  recent experiments have indicated that BP is efficiently
 converted  to mutagenic metabolites  by S-9 at < 0.1 mg/ml, followed by a
 decrease at higher  concentrations of S-9  protein.

      Determination of Expression Time.  Following induction of (genotypic)
 mutation,  a period, i.e., expression time,  is required to allow maximal
 phenotypic  expression  before drug selection.  In these experiments, the
 mutations were expressed during cell growth in suspension culture to
 minimize cell-to-cell  contact and to provide uniform growth conditions.
 The  CHD-AT3-2  subline  had a doubling time of 14 h  in suspension culture,
 and  the  spontaneous frequency of all four drug-resistant loci remained
 reasonably  constant (S.D. of 1 60%)  over  a 10 d period.   The relative
 survival for CHO-AT3-2 exposed to EMS is  shown in  Fig.  6.  Fig. 7 shows the
 relative survival (S)  of CHQ-AT3-2  cells  as a function of increasing
                                       30

-------
S-9 or microsome  concentration;  promutagens  coupled  with metabolic  activion
were DMN (9 mM) and BP  (0.04 mM).

    In the multiple-marker system, expression  of AAr mutants  induced  by
EMS was complete  by 2 d  (Fig. 8a).  The BP-induced AAr mutant  frequency
(Fig. 8b) appeared to be maximal by 4 d, declining slightly but  not signi-
ficantly over the next  6 d.  The DMN-induced AAr mutant phenotypes  (Fig.
8c) were maximally expressed by  3 d, and a plateau in induced  mutant
frequency was maintained through 10 d.  The  longer expression  time  needed
for complete expression  of TG resistance (compared to AG resistance)  has
been documented previously (28).  Mutants induced by EMS showed  full
TG-resistance by  6 d (Fig. 9a).  The frequency of TGr mutants  in response
to BP was also maximal  by 6 d, with no increase observed during  the
following 4 d (Fig. 9b).  The DMN-induced mutant frequency for TGr  shown
in Fig. 9c increased up  to 8 d and had plateaued by  10 d.

    In Fig. lOa, OUA^ mutants were fully expressed by 3 d and  remained
reasonably constant over the next 7 d.  With CHO-IB-2 cells, we  observed a
pronounced decline in the number of MNNG-induced OUA^ mutants  after 10-12
d (data not shown).  The expression of FUdRr mutants is shown  in Figs.
lOb and lOc.  Mutants induced by EMS were maximal at 2-3 d, then declined.
However, the response to BP declined after 1 d.  The frequency of mutants
induced by DMN was maximal at 3 d, declining significantly at  4  d and then
remaining constant over  the 10 d period.  Response at the ^k  locus  following
exposure to several mutagens has shown maximal expression at  1-4 d  (Adair
and Carver, in preparation).

    Multiple Marker Dose Response for Optimal Expression.  In  Figs, lla-lld,
the induced mutant frequency per viable cell is fitted as a linear  function
of the response to mutagen dose in units of  the logarithym of  decreasing
cell survival.  The slope of the linear function can be used  to  compare the
mutagenic potency at equivalent cell survival.  If the AAr response (Fig.
lla) after BP treatment  is compared to that  for DMN, i.e. a ratio of  the
slopes, then DMN  and BP are comparable mutagens at the aprt locus
(P >0.80).  In terms of equal cell survival, EMS is about seven times more
potent than DMN or BP.   In Fig. lib, the response of TGr mutants  induced
by BP is shown.  The response is slightly less than  that for DMN (P < 0.06
that the mutagens are comparable in potency).  Similar to AAr, the TGr
response for EMS  is about six fold greater than that for DMN or  BP.   In
Fig. lie, BP and DMN are similar in potency  (P~0.30) for inducing OUAR
mutants.  The response  for EMS at this locus is about seven times greater
than that of DMN  and BP.

    The above data have compared the mutant  frequency observed at optimal
(7-8 d) expression.  For the tk locus, the data are  compared in  Fig.  lid
for optimal but varying expression, i.e. 1 d for BP, 3d for DMN and EMS.
At early expression, DMN and BP are comparable mutagens; however, EMS is
slightly less effective  than DMN and BP at the tk locus.
                                      31

-------
                                   r
                                  f
t
                            EMS
                                         1
                                  456
                                    Expression, days
              8
10
                                                                         8a
Fig. 8   Induced AAr mutant frequency per viable cell (CHO-AT3-2) as a
function of expression time before selection with 80 ug/ml AA.  Error
bars are 1 SEM.  The AA  spontaneous frequency (+_ 1 S.D.) was
1.35 +_ 0.85 x 10  .  The mutagen concentrations and corresponding
fraction of surviving cells at time zero were as follows:
    (a)  EMS: 2.4 mM, 0.53 {-£-); 2.9 mM, 0.40 (	A	)•
         3.2 mM, 0.33 ( --A-- ); 3.4 mM 0.24 (—«—).
    (b)  BP:  0.04 mM, 0.2 mg/ml S-9, 0.90 (—O—);
         0.04 mM, 0.4 mg/ml S-9, 0.76 (	«	);
         0.04 mM, 0.6 mg/ml S-9, 0.69 (--•").
    (c)  DMN: 9 mM, 0.5 mg/ml S-9, 0.94 (—o—);
         9 mM, 1.0 mg/ml S-9, 0.80 (	O	);
         9 mM, 1.5 mg/ml S-9, 0.55 (—•--).
                                      32

-------
in

 o
 f
 S
 3
 u

 TJ
         BP
                                          I       I
      1      I
                                          I	I
I	I
                                  456

                                     Expression, days
              9      10
                                                                                8b

9
in
X
c
Q>
§• 6
1
i
fe
•o
0
C
1 1 1 1 1 1 1 1 1
DMN j
T, 	 ,- 	 1 	
* f *
/
/ IT
- ' j n
;/,r-r
//
''J\ \ S ? ^

)123456789
Expression, days



-

__c
n

r

K

                                                                                8c
                                        33

-------
                                  4      5
                                Expression, days
8
10
Fig. 9   Induced TGr mutant  frequency  per  viable cell (CHO-AT3-2) as a
function of expression  time  before  selection with 4 pg/ml AA.  All symbols
as in Fig. 8-  Spontaneous TGr  frequency was 2.52 _+ 1.46 x 10~5.
                                      34

-------
in
I,
    21
    18
    15
     9
•o
                     I       I        I       I
                     I        I
            BP
                                   4567
                                    Expression, days
     21
     18
  X
  fr 15
  g
  a
  •a
               I       I        I
             DMN
                      Q      FJ
               123456
                                      Expression, days
                                                                               9b
I       I        I       I        I
                      8       9      10
                                                                               9c
                                         35

-------
3U
40
in
o
x 30
o
c
0)
3
Sf 20
«^
4->
«2 J
3
E
1 1 l l i
H- T !
1 =
I \
i J

E(V
—

••
rf

III!
i 	 \
k 1
> -3


IS
—

•>
i

  O
  o
  •a
     3  -
2 -
1 -
                                      DMN, BP
                                   l
                                  456
                                   Expression, days
                                                         8
                                                                        lOa
Fig. 10   Induced FUdRr or OUAR mutant  frequency  per  viable cell
(CHO-AT3-2) as a function of expression time  before selection with 2 jj g/ml
FUdR or 3 mM QUA.  All symbols as  in Fig. 8.   Spontaneous  frequency was
2.44 + 0.81 x 10~6 for OUAR and 2.39 +  0.86 x 10"3 for  FUdRr.
                                      36

-------
234567
              Expression, days
             XTT              UNll\
             r------
       1   *	_i_
             *	$	
              456
               Expression, days
8
10
          lOc
                  37

-------
                           0.5
                   0.10
                                 Relative Survival
                                                               lla
Fig. lla   Increase  in induced frequency of AAr mutants (7-8 d
expression) with  decreasing  relative survival or fraction of surviving
CHO-AT3-2  cells (S)  after  treatment with DMN (—-A-.-), BP (-—•—), or
EMS (—•—).  Open  and  closed symbols represent different experiments.
The linear regression of  the straight line fit (y = DQ + b^ X, where
X =  log S ) has  the  following parameters:  (_+ 1 S.D.):
DMN:  bQ = -0.27 + 0.86  x  10~5,
BP:   bQ = -1.75 + 1.57  x  10~5,

EMS:  bQ = 1.72 + 1.62 x 10~4,
= 4.83
                                                 0.28 x 10~4;
                                                          "4
                                        = 5.04 + 0.92 x 10

                                       = 3.33 + 0.77 x 10~3.
                                      38

-------
                                   Relative Survival
                                                               0.10
                                                                   lib
Fig. lib  Increase  in induced frequency of TGr mutants  (7-8  d  expression)
with decreasing  relative survival or fraction of surviving CHO-AT3-2 cells
(S) after treatment with DMN, BP, or EMS.  All symbols  as in Fig.  lla.
Regression parameters are:
    DMN:  bQ = 0.26  +_ 0.92 x 10~5,  b
BP:
EMS:  b
b0 = -1.69
                        1.58 x 10
                1.55  +  1.95  x 10
                                ~4
 = 7.66 + 0.35 x 10'4;

L = 5.45 + 0.85 x 10~4;
 = 4.43 + 0.93 x 10"3.
                                       39

-------
             1.00
                 0.50
                                                            0.10
                                 Relative survival
                                                      lie
Fig. lie   Increase  in induced frequency of OUA^ mutants  (7-8 d
expression) with  decreasing relative survival  or  fraction of surviving
CHD-AT3-2  cells (S)  after treatment with DMN,  BP,  or  EMS.  All symbols as
in Fig. lla.   Regression parameters are:
    DMN:

    BP:

    EMS:
bQ =  0.22 + 2.1 x  10
                                "6
                               1.63 + 0.12 x 10
                                               -4.
b0 =  0.12 + 3.22  x 10~6,  \>l = 1.42 + 0.17 x 10~4;
x 10"5, bj =
                                      °'25
                                             10
                                       40

-------
               900 -
             X

             fr
                600 -
             1
             cc
             3
                300
                 1.00
               0.50
                                                            0.10
                                     Relative survival
                                         lid
Fig. lid   Increase  in induced frequency of  FUdRr  mutants after optimal
expression time  with decreasing relative survival or fraction of surviving
CHD-AT3-2  cells  (S)  after treatment with DMN,  BP, or EMS.  All symbols as
in Fig. lla.  Regression parameters are:
DMN (3 d):
b0 = 0.47 ± 0.34  x
                                             =  1>93    0.26 x 1Q-2.
                      ~3
BP  (1 d):  bQ =  0.47  + 0.34 x 10,

EMS (3 d):  bQ =  0.14  + 0.07 x 10~3,b1
                             = 1.93
                                                       0.26 x 10~2;
                                               1.38  +  0.20 x 10
                                                               ~2
                                       41

-------
    Validation of Mutant  Phenotypes.  As  shown  in Fig.  12a,  of the 61  AAr
clonal isolates assayed for APRT  activity,  96%  (22/23)  of  the  spontaneous
mutants and 82% (31/38) of the mutants  from mutagenized cultures  had < 5%
wild type activity; only  4 AAr clones had APRT  activity approaching those
of the heterozygous parental cells.  Of the 54  AGr clones  assayed for
HGPRT activity, 89% (16/18) of the spontaneous  mutants  and 86% (31/36) of
the mutants from mutagenized cultures had  < 30% wild  type  activity; 4  AGr
clones had HGPRT activities similar to  those of CHO-IB-2 and CHO-AT3-2.
The mean residual HGPRT activity  of TGT mutants was much lower than that
of AGr mutants.  None of  the 30 TGr mutant  clones  assayed  had  WT  HGPRT
activity, with 94% (17/18) of the spontaneous mutants and  67%  (8/12) of the
induced mutants having <  5% WT activity.

    In Fig. 12b, the residual TK  activity of spontaneous and induced
mutants is shown.  Because the spontaneous  frequency  of FUdRr  is  high, no
attempt was made to distinguish between spontaneous and induced mutants.
Both classes had virtually no detectable TK activity; a single isolated
mutant appeared to have 15% of WT TK activity,  but  the  line  was lost before
the assay could be repeated.  All other mutants have  had <  1% WT activity.
After treatment with DMN  or BP, fifty induced mutant  clones  resistant  to
2 y g/ml FUdR were isolated.  The  clones were grown  12-16 generations in the
absence of FUdR, then tested for FUdRr, BUdRr,  TFTr,  and HAT8.  All
fifty clones displayed cross-resistance to  the  three  thymidine analogs
(P.E. 0.40-0.70) and showed complete sensitivity  to HAT (P.E.  < 10"^).

-------
              30
3j 20
ID

I
I 10
                      I • •  ' ' I ' • • '  I ' • • ' I

                        Spontaneous AG' LI

                                TG' 0
                                     I i ' '  ' I ' ' ' '  I ' ' ' ' I

                                         Induced AG' 0

                                               TG' 0
                                                      • i i i n m i i n~
              30
S
              20
            g
            o
            O 10
                        Spontaneous AA'
                      25    50    75    100   0    25    50    75    100
                      % Wild type activity               % Wild type activity

48
0)
(Q
y,
(9
S
§24
ii i i | i i i i | i i i i | i i i
Spontaneous and induced
™
-

£
V

FUdRr D

BUdRr 0
TFTr •
o m
1 1
-
^~
-


-
—
                                 25      50     75
                                  % Wild type activity
                                           100
Fig. 12   Activities of  HGPRT, APRT, or  TK (relative to wild type) for
spontaneous  and induced mutants.  Mutant  clones were  isolated,  grown 10-12
generations  in the absence  of selecting drug,  and assayed  for in vitro
activity  of  the appropriate enzyme,   (a)  Mutants resistant to 40 Mg/ml AG
or 4 ug/ml TG or 80 pg/ml AA; (b) mutants resistant to 2 ug/ml  FUdR,
150 yg/ml BUdR, and 3 ug/ml TFT.
                                        43

-------
                                   SECTION 6
                                  DISCUSSION

    As  an  in  vitro  test  system,  CHO  cells combine  a near  diploid karyotype
 that  facilitates  genetic  and  cytogenetic  analysis  with  the  favorable  growth
 characteristics,  e.g., high plating  efficiency  and rapid  growth in either
monolayer  or  suspension  culture,  necessary for  large-scale,  cost-effective
mutagen  screening.   The  CHO-AT3-2 subline, which is also  heterozygous at
 both  the aprt and tk loci, was  developed  to  allow  single-step selection of
 autosomal  recessive  AAr  and FUdRr mutant  phenotypes,  as well  as the
more  commonly used X-linked TGr  and  co-dominant OUAR genetic  markers.
Although functionally heterozygous aprf1"'  strains of CHO had been
 isolated previously  (17,52),  optimal conditions for the phenotypic
expression and selection  of AAr mutants were  not established,  and  the
 strains  were  not  extensively  used for quantitative mutagenesis  assays.
Only  limited  data have been reported in the  literature  regarding the
optimizations  of  selection parameters and the induction of  APRT-deficient
mutants  by known  mutagens (17,34,52,87,92).   Presumptive  J:k+/~  revertants
of TK-deficient mutants  have  been isolated in several cell  lines
 (3,10,22,80,85).  Extensive validation and application  of the L5178Y
_tk~/~ mouse lymphoma system for mutagenesis  assays (20,22,24,25) prompted
our efforts to use the tk locus  in a CHO  cell mutagenesis assay system.
The CHO-AT3-2  subline facilitates  simultaneous, quantitative  mutagenicity
testing  at four specific gene loci.

    Derivation of heterozygous cell  lines.   Phenotypic  expression  of
single-step forward mutations at  the autosomal aprt and tk  loci,  and  use of
the AAr  and FUdRr genetic markers  for mutagenesis  assay,  first  required
the derivation of a  CHO subline  that was  heterozygous for these loci.
Using the approach of Jones and Sargent (52), an aprt +/~ heterozygote
with approximately 50% wild-type  APRT activity  (CHO-IB-2) was selected  from
an unmutagenized CHO cell population.  The aprt +/~ CHO-IB-2  cells were
subsequently used to derive sublines that were also heterozygous for  the tk
locus.   Since direct, single-step  selection of _tk  +'~ heterozygotes
proved to be  impractical (2,22,67), such  heterozygotes were obtained  by
first isolating rare, BUdRr,  ttc ~/~ mutants of CHO-IB-2,  and
subsequently  selecting HATr revertants which  had intermediate levels  of
TK activity.  A similar approach was used by Clive  and co-workers  (22)  to
obtain the tk +/~ heterozygote employed in their L5178Y TK +'~  mouse
lymphoma mutagenesis assay system.

    Biochemical and  genetic characterization.  The  CHO-AT3-2  cells are
heterozygous for both the aprt and ^k loci.  Consistent with  this  premise


                                     44

-------
were:   (a)  the manner  in  which  the  CHO-AT3-2  cell  line  was  derived (Fig.  1
&  2);  (b) this line's  intermediate  levels  of  APRT  and TK activity (Fig.  1 &
2);  (c)  the high  frequencies  at  which APRT- or  TK-deficient mutants  were
obtained from unmutagenized CHO-AT3-2 cell populations  in single-step
selections  with AA  or  FUdR (Table 1, Fig.  3); and  (d) forward  mutation
rates  at these autosomal  gene loci  that  were  greater  than or similar to
those  at the X-linked  hgprt locus (Tables  2 & 3).   Spontaneous and
EMS-induced frequencies of AAr  and  BUdRr or FUdRr  mutants were at
least  2-3 orders  of magnitude higher in  CHO-AT3-2  cells  than in wild-type,
aprt +/+ and tk +/+ CHO cells (Table 1).   Segregation analysis (Table 5)
confirmed that CHO-AT3-2  cells were indeed functionally  hemizygous for the
genes  for APRT and TK.

    Biochemical validation of drug resistant  mutant phenotypes revealed
that fifty-one of fifty-nine  clones selected  on  the basis of their
resistance  to AA  had < 5% of  wild-type APRT activity.  APRT-deficient
mutants have been selected from  several  other cells lines (17,33,52,87,-
92,93), and the aprt locus has been mapped in both the mouse (59) and human
(93) karyotypes.  Functionally heterozygous,  aprt  +'~ strains  have not
been previously utilized  for  quantitative mutagenesis assay,  and only
limited data has  been  reported in the literature regarding  the induction  of
AA-resistance by  known mutagens  (15,16,52,92).  Dose-dependent increases  in
the frequency of AAr colonies (~10 fold  over  background)  after treatment
of an  aprt +'~ strain  of CHO  cells with  EMS were reported by Jones and
Sargent (52), but expression  kinetics for  the AAr  marker  were  not
determined.  Having established  optimal  conditions  for the  phenotypic
expression and selection of AAr  mutants  in our  system (15,16),  we can
directly compare  induced mutation frequencies at the  aprt and  hgprt  loci,
and assess  the relative sensitivity of the genes for  APRT and  HGPRT  to
mutation by various physical  and chemical agents.   For the  direct-acting
mutagen, EMS, induced mutation frequencies for AA-resistance were only
slightly lower than those for TG-resistance (Fig.  3).  However,  we have
observed mutagen-specific differences in the  relative mutability of  the
aprt and hgprt loci using other  mutagens (15,16; Adair &  Carver,  manuscript
in preparation).

    Although TK-deficient strains have been isolated  from a number of
different cell lines (10,22,55-57,80,85,86), many  of  these  variants  were
obtained by prolonged, multi-step selection with BUdR, and  are of uncertain
genetic origin.  The presumptive structural gene locus for  TK  has been
mapped in both the mouse  (60) and human  (36,66) karyotypes,  using
TK-deficient mutants of mouse and Chinese hamster  cells.  Derivation of a
^k *'  heterozygote of L5178Y mouse lymphoma  cells  by Clive and
co-workers (22),  and extensive validation and application of this  system
for autagenesis assay  (21-23,25,26,67),  has prompted  efforts to  utilize the
tk locus in other mammalian cell mutagenesis  assay  systems  as  well (16,85).

    We have isolated TK-deficient mutants of CHO cells using BUdR, FUdR,  or
TFT as the selecting agent.  All but one of forty-three  thymidine
analog-resistant clones isolated from CHO-AT3-2 or CHO-IB-2 had  < 1% of
wild-type TK activity.  All of the TK-deficient mutants we  have  tested were
HAT-sensitive and cross-resistant to BUdR, FUdR, and  TFT.   Further
                                      45

-------
 biochemical  and  genetic  characterization of several TK-deficient mutants
 and  HATr  revertants  with intermediate  levels of TK activity provided
 evidence  that  they probably  represent  mutations in the  structural gene for
 TK  (Adair, manuscript  in preparation).   The thymidine  analog-resistant,
 TK-deficient mutant  phenotypes  were  recessive,  and complementation analysis
 (Table  4)  indicated  that all  of the  BUdRr,  FUdRr,  or TFTr  clones
 tested  represented the same  complementation group, reflecting mutation at
 the  same  genetic  locus.   Furthermore,  non-complementation  of the thymidine
 analog-resistant  phenotype in hybrids  with  TK-deficient DON-a3 cells
 identified this  genetic  locus as  the presumptive structural gene locus for
 TK,  previously mapped  in both the mouse  and human  genomes  using the DON-a3
 strain  (36,60).

     Most  previous studies have  utilized  BUdR as the selecting agent for
 TK-deficient mutants (10,21-23,25,26,40,55-57,80).   The killing of
 TK-positive wild-type  cells  by  this  drug is rather slow, and several cell
 divisions  can  occur before cell death  (2,42,67,94).  FUdR  rapidly inhibits
 cell division  (42,94), effectively eliminating  background  growth of
 wild-type  cells; unlike  BUdR, it  is  apparently  not mutagenic in mammalian
 cells (50).  Both analogs are readily  phosphorylated by TK,  but FUdR is  not
 further metabolized  to the di-  or tri-phosphates,  or incorporated into DNA
 (42,43).  FUdMP  is a potent,  irreversible inhibitor of  thymidylate
 synthetase (43,79).  The observed cross-resistance of TK-deficient mutants
 to BUdR, FUdR, and TFT,  along with the non-complementation  of TK-deficient
mutants regardless of which drug was used for their selection (Table 4)  and
 the  similar rates of spontaneous mutation for resistance to each drug in
 CHO-AT3-2  (Table 3), together suggest  that  BUdR and FUdR select for the
 same class of  TK-deficient mutant phenotypes.   Therefore, we have used FUdR
 as the selecting agent for such mutants  in  these CHO-AT3-2  multiple-marker
mutagenesis experiments.

    The ^k locus appears  to be  considerably more susceptible to spontaneous
mutation than  the genes  for APRT, HGPRT, or  Na+/K+ ATPase  (OUAR)  in
Chinese hamster cells.   Background frequencies  of  BUdRr or  FUdRr
 colonies for unmutagenized CHO-AT3-2 cell populations were  consistently
 1.5 - 3.0 x 10~-\  Similar spontaneous frequencies  have been reported for
BUdRr variants in several presumptive ^k +'~ heterozygotes  of another
Chinese hamster cell line (80).  Luria-Delbruck fluctuation  analysis
 (Table 3) indicated that  the  high frequencies of thymidine  analog-resistant
variants in CHO-AT3-2 cell populations reflect  a very high  spontaneous rate
of mutation at the j:k locus.  Stable,  thymidine  analog-resistant  variants
arise spontaneously at a rate of 2-3 x 10~^  per  cell per generation;  this
rate is approximately three orders of magnitude  greater than  those  for
spontaneous mutation at  either  the aprt  or  hgprt loci in the  same  cell line
 (Table 2).  Despite the  high  background  frequencies  of  FUdRr  colonies,
 significant dose-dependent increases in mutation frequency  were  observed
with this genetic marker after  treatment with known mutagens  (Fig.  3).
Furthermore,  for most of the mutagens we have tested in the  CHO-AT3-2
system, more mutations appear to be induced  at  the £k locus  as  a  function
of mutagen dose than at  either  the aprt  or  hgprt loci,  or in  the  genes
involved in ouabain resistance  (16;  Adair & Carver, manuscript  in
preparation).  Clive e_t £l_.  (26) reported that  the ^k locus was,  on  the


                                     46

-------
average, nearly  10  times more mutable  than was  the  hgprt  locus,  for  13
agents  tested  in the L5178Y TK +/~ mouse  lymphoma system.   Several
carcinogens  that were  only weakly or non-mutagenic  at  the  hgprt  locus were
found to be  strongly mutagenic at the ^k  locus.  Clive  (26)  has  suggested
that the L5178Y  system, utilizing the  tk  locus, might  be  allowing  the
detection of TK-deficient mutants with  gross  chromosomal  lesions as  well as
those generated  by  base-pair substitutions or small  deletions.  The
extraordinarily  high rate of spontaneous  mutation at the ^k  locus  in
CHO-AT3-2 cells  may reflect a position  effect or "hot  spot"  for mutations
or chromosomal events  such as mitotic recombination, segregation,  or
non-disjunction.  Further investigation of this system may  provide insight
into the involvement of extra-mutational  events in  the expression  of
genetic damage in somatic cells.

    Cell density parameters.  A major  technical consideration  in mutagen
testing assays is the  dependence of mutant selection upon wild type  cell
density, i.e., fewer mutants can survive  in high density  cultures  because
of metabolic cooperation (contact feeding).  Cell density effects  have been
previously quantitated for several drug resistance markers  in  other  cell
lines (32,34,52,67,71,78,80,88,92).  However, to establish optimal cell
plating density  conditions for the selection of AAr, AGr, TGr,
OUA^, and FUdRr  mutants in our cell line, the mean  critical  density
MCD50 (34),  of CHO-IB-2 and CHO-AT3-2 cells for each drug resistant
phenotype was  determined.

    Reconstruction  experiments (Fig. 4),  and determinations  of spontaneous
mutant frequencies as  a function of cell  plating density  (Fig. 5),
indicated that critical cell densities  for TG, AG,  and AA were 7.5 x 103
cells per cm2, 3.0 x 10^ cells per cm2, and 1.3 x 10^ cells  per
cm2, respectively.  Our cell density data for AA resistance  are very
similar to those  reported in a detailed study of CHO-K1 by Dickerman and
Tischfield (34)  and do not show the density independence of  AAr
previously reported in another CHO cell line (52).  Critical densities
ranging from 4 x 102 to 4 x 10^ cells per cm2 for AG-resistance
(32,41,51,71,78,  97),  and from 6 x 103  to 1.4 x 106 cells per  cm2 has
been reported  for TG-  resistance (27,31, Corsaro and Migeon  as calculated
in 34,58,71).  The selection of ^k~'~ phenotypes was not dependent upon
cell density over the  range of 2 to 20  x  103 cells  per cm2.  In
general, cell  lines plated in semi-solid  agar substrates, e.g. L5178Y, S49,
are not thought  to show cell-to-cell cooperation in mutant selections
(27,41).  However, a recent report (67),  indicated  that BUdR (but not TFT)
selection of L5178Y mutant cells in agar  showed decreased recovery at 4 x
10^ cells per  cm2.

    Concentration of Selecting Drug.  Cumulative data suggest  that AG must
be used at concentrations > 30 ug/ml to select reproducibly  bona fide
hgprt" mutants (12,39,41,53,78).  Since selection of TGr mutants is
affected by metabolic  cooperation at densities approximately 4-5 fold lower
than AGr (Figs.  4, 5a), we initially attempted to optimize the AGr
selection and obtain quantitative, reproducible data.  Although mutant
selection can be improved by using 30 to 40 ug/ml AG, selection of hgprt"
mutants with the  CHO-IB-2 cell line is more definitive and quantitatively
                                      47

-------
 reproducible  when  TG  (2-4  ug/ml)  is  used  as  the  selecting agent,  an
 observation consistent  with  results  reported in  other  cell  lines
 (47,58,88,95).   Furthermore,  as shown  in  Fig.  12,  the  spectrum of residual
 HGPRT  activity  is  much  broader for mutants selected  by resistance to  AG;
 the  resulting heterogeneity  in colony  size makes  quantitation  more
 difficult  and requires  extended incubation of  mutant colonies  (up to  18  d).
 Therefore, we prefer  TG for  selecting  HGPRT-def icient  mutants  in  our
 system.  Azaadenine at  40  to  80 ug/ml  effectively  selects APRT-deficient
 mutants.  While  BUdR  used  at  150 ug/ml  avoids  most of  the background  haze
 associated with  abortive phenotypes, the  quantitation  of  TK-def icient
 mutants is more  satisfactory  with FUdR at 1  to 3 ug/ml; very definitive
 data are also obtained  with TFT at 1 to 3
    Expression Time.  Optimization of  expression  time  is  a major  factor  in
the reproducible selection of drug-resistant mutants (48,84,89).  Although
expression  time is generally expressed either  as  days  or  population
doublings (or cell generations), both  chronological time  and  cell division
seem  to contribute to expression (32,48,54).   The HGPRT activity  in  a  new
presumptive mutant (and consequently its potential sensitivity  to cytotoxic
selecting agents) decreases as a function of both degradation of  preexisting
enzyme and  dilution by cell division (48).  Althogh HGPRT is  apearently
degraded even in contact-inhibited, nondividing cell populations  (48),
reports also indicate that maximal mutation frequencies are not achieved
when  cells  are maintained in a nondividing state  (54).  Cox and Masson (32)
presented evidence suggesting that cell doubline parameters may be more
important than chronological time, at  least in the expression of
radiation-induced mutation at the hgprt locus.  With direct,  in situ
expression  of mutagenized cell populations, prolonged  expression  times are
not practical because of cell-to-cell  density  effects  and metabolic
cooperation.  With the replating method, the optimal expression time
generally extends over at least several generations (1,13,32,48,69,74,96).

    In CHO-AT3-2 cells, the optimal expression time depends on both  the
drug-resistant marker and the mutagen used (Figs. 8,9,10).  Mutagen-
specific variation in expression time has also been reported  for diploid
human lymphoblasts (91), in which MNNG-induced TG resistance  was expressed
by 6 generations but cells treated with 1CR-191 required  12 to 16
generations before full expression had occurred.  Shapiro and co-workers
(82) reported maximal expression of TGr mutants after  2 to 6  generations,
but expression times from 2 to 20 d have been  reported by other
investigators (1,27,32,48,58,74,82,89,91,96).  In CHO-AT3-2,  6 to 8 d  (8 to
10 generations) were required for maximal expression of TGr mutants
(Fig.  9).

    Optimal expression times for AG-resistance ranging from 3 to 7 d have
been reported (32,39,51,69).  Khalizev, e± a±. (54) reported  that 2 to 3
doublings  were sufficient for maximal expression of the AGr mutants
phenotypes in Chinese hamster cells,  but our data for CHO-IB-2 indicated
that 5 to 7 doublings were required for optimal expression of AGr mutants
induced by EMS or MNNG (data not shown).
                                      48

-------
     In CHO-AT3-2  cells, 2  to 4 d  are needed  for  full  expression  of  AAr
phenotypes  (Fig.  8).  For  QUA resistance, maximal  expression  reportedly
occurred in 1  to  4 d  (27), while  in our system 3 d (3 to 4 doublings) were
sufficient  for expression  of most OUAR mutants (Fig.  lOa).  The  mutant
frequency was  constant over the 10 d, whereas Buchwald (9) observed a
decrease in the frequency  of OUAR mutants after  2  generations of
expression.  For  the ^k locus, the optimal conditions for mutant expression
are  complex and. mutagen-specific.  For the mutagens studied,  the maximal
observed mutant frequency  varies  from 1 to 3 d (Figs.  lOb, lOc).  Previous
reports indicated 2 to 3 d for mouse lymphoma (26), 3 d for V79  cells (80),
and  a constant 2  doublings (generations) for BHK cells treated with EMS
(10).

     Optimal  expression time clearly varies for different mutagens,  and the
time for maximal  expression observed for a few standard mutagens like EMS
may  be of limited value in predicting the optimal  expression  time for other
classes of mutagens or for complex mixtures.  If quantitative comparisons
are  to be made between mutagens,  optimal expression time may need to be
established  for each mutagen, e.g., BP vs. DMN at  the tk locus.

     Dose Response Relationships at Different Genetic  Markers.  When the
observed frequency of AAr, AGr, TGr, and OUAK mutants  per viab1e
cell is compared  for CHO-IB-2 as  a function of molar  dose (Table 6), both
NQO  and MNNG are more effective than EMS (approximately 200 to 350  fold and
100  to 500  fold,  respectively).   However, when the induced frequency of
mutants per viable cell is compared for the various markers as a function
of log cell  survival, NQO  and EMS are 2 to 7 and 3 to 9 fold more mutagenic
than MNNG (Table  6).  Thus, as previously suggested (29,77), when analyzed
on the basis of cell survival, MNNG is not a particularly effective  mutagen
in mammalian cells.

    As a function of applied molar dose, the mutational response at  the
hgprt locus  in CHO-AT3-2 cells had a slope of 5.2  x 10"^ mutants per
viable cell per mM EMS, compared  to a slope of 4.3 x  10~^ previously
reported for this locus by Hsie and co-workers (45) at similar 16 h
treatment times.  Slopes for AAr, OUAR, and FUdRr mutants after  8 d
expression were 4.1 x 10~^, 1.3 x 10"^, and 1.3 x  10"-*, respectively
(3).

    When the mutational response  for EMS is plotted as a function of cell
survival, the  observed frequency  of AAr, TGr, and OUAR, mutants  by
EMS decreases  at  survivals   0.50 (Fig. 11), as was the case in  previous
experiments using CHO-IB-2 cells  (Table 6).  However,  induction  of  FUdRr
mutants by EMS increased linearly as a function of the logarithym of
decreasing  survival, as did induction of AAr or TGr mutants by DMN.

    Specificity and sensitivity of the multiple markers.  Significant
increases in mutation frequency were observed for all  four drug  resistance
markers following treatment of CHO-AT3-2 cells with EMS, DMN, or BP  (Fig.
11), but locus-specific differences in sensitivity to  these mutagens were
noted.  Under  optimal expression  conditions, the tk locus consistently
showed the greatest mutational response as a function  of cell survival with
                                      49

-------
 induction  by  DMN = BP  > EMS.   The  jtk locus was much more sensitive than the
 other  loci  to mutation induction by DMN and BP, but only slightly more
 sensitive  to  mutation  induction by EMS.  With DMN and BP, the slopes for
 mutation induction for FUdRr  (Fig. 11)  were about 25 to 40 fold greater
 than  those  for  AAr or  TGr,  and over 100 fold greater than those for
 OUAR.  With EMS,  mutation  induction slopes for FUdRr were still about 3
 to 4  fold  higher  than  those  for AAr or  TGr, and about 12 fold higher
 than  those  for  OUAR.   EMS,  however,  was much (7 to 9 fold) more effective
 in inducing AAr,  TGr,  and OUAR mutants  than either DMN or BP.

    Of the  four genetic markers employed in CHO-AT3-2 mutagenesis assay
 system, the tk  locus has consistently  shown the greatest mutational
 response to positive mutagens  (Ref.  3,  Fig. 11).   The magnitude of induced
 mutation at this  locus suggest that FUdR selection in the CHO-AT3-2 system
 may detect  a  broader range  of  genetic  lesions than the other  drug-resistance
 markers.   In  the  L5178Y TK"4"'   mouse lymphoma system,  the ^k locus appears
 to be approximately  ten times  more mutable than the hgprt locus (26),  and
 Clive and  co-workers have postulated that their system may be detecting
 TK-deficient  variants  arising  from gross chromosomal  lesions  as well as
 those generated by base-pair substitution,  frame  shift,  or small deletions.
 In our CHO-AT3-2  cells,  as  in  the  L5178Y TK+/~ system,  phenotypic
 expression  of mutants  resistant to thymidine analogs  was rapid, with
 maximal frequencies generally  attained  within 1 to 3  d after  mutagen
 treatment  (Fig.  10).   Mutant frequencies typically decline with longer
 expresstion times.  In both systems,  cell isolates resistant  to thymidine
 analogs have  little or no detectable TK activity  (22,25) as would be
 expected in the case of deletions  or gross  chromosomal lesions.

    The rapid expression kinetics  for  induced FUdRr mutants may
 facilitate  the  detection of mutants  with gross chromosomal anomalies or
 large scale DNA damage affecting cell growth rate.   Such gross  genetic
 lesions might not  be detected  using  either  the OUAR marker (a functional
 Na+-K+-ATPase is  required for  cell survival)  or the TGr  marker
 (expression of  TGr phenotypes  requires  long expression times).   Moreover,
 the gene for  _tk may be unusually susceptible  to mutation.   The  high
 frequencies of  FUdRr observed  in unmutagenized CHO-AT3-2 cell populations
 are the result  of  an exceptionally high,  apparent  rate  of spontaneous
mutation at the ^k locus (3).   The apparent susceptibility to mutation of
 the £k locus  in CHO-AT3-2 cells may  reflect chromosomal  events  such as
mitotic recombination,  segregation,  or  nondisjunction.   Position effects in
 lower eucaryotes  and "hot spots" in  microbial  systems  are not uncommon.
 Since the FUdRr marker may allow detection  of a broader  range of genetic
 lesions than  the  other drug resistance  markers, use of  the tk locus in the
 CHD-AT3-2 mutagenesis  assay enhances the potential sensitivity  of the
 system.

    Validation  of mutant phenotypes.  As shown in  Fig.  12,  the  spectrum of
 residual enzyme activities of  AGr  mutants  is  more  heterogeneous than that
 for TGr mutants.  Although many of the  mutants selected  with  AG display
 stable characteristics,  the observed heterogeneity probably contributes  to
 technical problems with  the selection of AGr  phenotypes.   There was no
                                      50

-------
statistical difference  in  the  residual  enzyme  activities  in EMS-,  MNNG-,
and NQO-induced mutants  (data  not  shown).  Mutants  selected with FUdR,
BUdR, and TFT appear  to  have identical  enzymatic  properties,  displaying
<  1% WT TK activity,  sensitivity  to HAT,  and cross  resistance to all  three
selecting drugs.  The results  in Fig.  12  indicate  that  a  very high
percentage of the spontaneous  and  induced  drug-resistant  clones scored  as
mutants have substantially diminished  enzyme activity at  the  selected locus.

    The Multiple Marker  Mutagenesis Assay  System.   Systematic determination
of optimal conditions for  the  phenotypic  expression and selection  of  AAr,
FUdRr, OUAR and TGr mutants (15,16; Adair  & Carver,  manuscript in
preparation), and the development  and  application  of a multiple-marker
mutagenesis assay system employing CHO-AT3-2 will  facilitate  quantitation
and direct comparison of the extent of mutational  damage  at  these  four
defined genetic loci.  The CHO-AT3-2 cells combine  favorable  growth
characteristics (e.g., short doubling  time, excellent cloning efficiency,
and the ability to grow  in either monolayer or suspension culture) with a
near diploid genome that facilitates genetic and  cytogenetic  analysis.  As
an in vitro mutagenesis  assay  syste, the CHO-AT3-2  cell line  offers the
unique capability of measuring mutation induction  at the  autosomal aprt and
tk loci as well as in the genes involved  in ouabain resistance or  at  the
X-linked hgprt locus.  The assay thus  combines the  attributes of the  two
major in vitro mammalian-cell  mutagenesis  systems  currently  being  used
Furthermore, it offers the only CHO assay  presently using the J:k locus  and
facilitates comparison with results from  the L5178Y mouse lymphoma system
(20,22,24,25,26).  Such  comparisons will be invaluable in determining
species variation among  short-term, in vitro mutagenesis  assays.   To
develop this system and  confirm its potential, we  have:   (a)  selected CHO
stock lines heterozygous at the aprt and  the ^k loci (necessary for
single-step selection of autosomal recessive mutations),  (b)  tested stock
cell lines for optimal plating density, expression  time,  and  concentration
of selecting drugs, (c)  validated the  assay with  data for induced mutant
frequencies after treatment with the direct mutagens EMS, MNNG, and NQO and
promutagens DMN and BP,  (d) quantified and defined  the dose  response, and
(e) confirmed the mutational events by assays  for  enzyme  activities in
mutant clones isolated from the above  experiments.

    Because the mechanisms by  which mutational damage is  incurred may vary
considerably for different agents, and genetic loci  may differ in  their
sensitivity to those agents, systems employing multiple markers may be  more
effective in detecting potentially mutagenic agents.  The operating
parameters for our multiple marker mutagenesis assay are  now  well-defined
and the system can be applied  to testing other known mutagen  classes  as
well as to screening samples containing mixtures  of  potentially mutagenic
agents.  Coupled with the metabolic activation methods, the  system can  be
used to screen crude mixtures  and effluents of environmental  concern.
                                      51

-------
                                  References
 1.    Abbondandolo, A., Stefania Bonatti, C. Colella, G. Corti, F.
       Matteucci, A. Mazzaccaro, and G. Rainaldi, A comparative study  of
       different experimental protocols for mutagenesis assays with  the
       8-azaguanine resistance system in cultured Chinese hamster cells,
       Mutation Res., 37 (1976) 293-306.

 2.    Adair, G.M. and J.H. Carver, Unstable, non-mutational expression of
       resistance to the thymidine analogue, trifluorothymidine in CHO
       cells, Mutation Res., 60 U979) 207-213.

 3.    Adair, G.M., J.H. Carver and D.L. Wandres, Mutagenicity testing in
       mammalian cells.  I.  Derivation of a Chinese hamster ovary cell
       line heterozygous for the adenine phosphoribosyltransferase and
       thymidine kinase loci, submitted to Mutation Res.

 4.    Adair, G.M., L.H. Thompson, and S. Fong,  -^H  Amino acid selection
       of aminoacyl-tRNA synthetase mutants of CHO cells:  Evidence of
       homo- vs. hemizygosity at specific loci, Somat. Cell Genet. 5 (1979)
       329-344.

 5.    Arlett, C.F., D. Turnbull,  S.A. Harcourt, A.R. Lehmann, and C.M.
       Colella, A comparison of the 8-azaguanine and ouabain-resistance
       systems for the selection of induced mutant Chinese hamster cells,
       Mutation Res., 33 (1975) 261-278.

 6.    Baker, R.M., D.M. Brunette, R. Mankovitz, L.H. Thompson, G.F.
       Whitmore, L. Simonvitch, and J.E. Till, Ouabain-resistant mutants of
       mouse and hamster cells in culture,  Cell, 1 (1974) 9-21.

7.      Boag, J.W., The statistical treatment of cell survival data,  in: T.
       Alper (Ed.), Cell Survival  after Low Doses of Radiation: Theoretical
       and Clinical Implications,  The Instutute of Physics and John Wiley
       and Sons, Bristol,  1975, pp. 40-53.

 8.    Bridges, B.A., The  three-tier approach to mutagenicity screening and
       the concept of radiation-equivalent  dose, Mutation Res., 26 U974)
       335-340.

 9.    Buchwald, M.,  Mutagenesis  at the ouabain-resistance locus in  human
       diploid fibroblasts, Mutat. Res.,  44 (1977)  401-412.


                                      52

-------
10.    Caboche, M., Comparison of the frequencies of spontaneous and
       chemically induced 5-bromodeoxyuridine-resistance mutations in
       wild-type and revertant BHK-21/13 cells, Genetics, 77 U974) 309-322.

11.    Capizzi, R.L. and J.W. Jameson, A table for the estimation of the
       spontaneous mutation rate of cells in culture, Mutation Res., 17
       (1973) 147-148.

12.    Carson, M.P., D. Vernick and J. Morrow, Clones of Chinese hamster
       cells cultivated in vitro not permanently resistant to azaguanine,
       Mutation Res., 24 (1974) 47-54.

13.    Carver, J.H., W.C. Dewey and L.E. Hopwood, X-ray-induced mutants
       resistant to 8-azaguanine.  I. Effects of cell density and
       expression time, Mutation Res., 34 (1976) 447-464.

14.    Carver, J.H., W.C. Dewey and L.E. Hopwood, X-ray-induced mutants
       resistant to 9-azaguanine.  II. Cell cycle dose response, Mutation

15.    Carver, J.H., D.L. Wandres, G.M. Adair, and E.W. Branscomb,
       Mutagenicity testing in mammalian cells:  The development of
       multiple drug-resistance markers having practical application for
       screening potential mutagens, Mutation Res.,  53 (1978) 96-97.

16.    Carver, J.H., G.M. Adair, and D.L. Wandres, Mutagenicity testing in
       mammalian cells. II. Validation of multiple drug-resistance markers
       having practical application for screening potential mutagens,
       submitted to Mutation Res.

17.    Chasin, L.A, Mutations affecting adenine phosphoribosyl transferase
       activity in Chinese hamster cells, Cell, 2 (1974) 37-41. Res., 34
       (1976) 465-480.

18.    Claxton, L.D. and P.Z. Barry, Chemical mutagenesis: An emerging
       issue for public health, Amer. J. of Public Health, 67 (.1977)
       1037-1042.

19.    Cleaver, J.E., Induction of thioguanine- and  ouabain-resistant
       mutants and single-strand breaks in the DNA of Chinese hamster ovary
       cells by 3H-thymidine, Genetics, 87 (1977) 129-138.

20.    Clive, D., A linear relationship between tumorigenic potency in vivo
       and mutagenic potency at the heterozygous thymidine kinase (.TIC*')
       locus of L5178Y mouse lymphoma cells coupled  with mammalian
       metabolism, 1977 Elsvier/North-Holland Biomedical Press, Progress in
       Genetic Toxicology, 1977, pp. 241-247.

21.    Clive, D, Recent developments with the L5178Y TK heterozygote
       mutagen assay system, Environ. Health Perspect., 6 (1973) 119-126.
                                      53

-------
22.    Clive, D., Flamm, W.G., Machesko, M.R.,  and  Bernheim,  N.J,   A
       mutational assay system using  the thymidine  kinase  locus  in  mouse
       lymphoma cells, Mutation Res.,  16 (1972)  77-87.

23.    Clive, D., G. Flamm,  and J.B.  Patterson,  Specific-locus mutational
       assay systems for mouse lymphoma cells,  in:  A. Hollaender (Ed.;,
       Chemical Mutagens, Principles  and Methods for  their Detection, Vol.
       3, Plenum Press, New  York,  1973, pp.  79-103.

24.    Clive, D. and J.F.S.  Spector,  Laboratory  procedure  for assessing
       specific locus mutations at  the TK locus  in  cultured L5178Y  mouse
       lymphoma cells, Mutation Res., 31 (1975)  17-29.

25.    Clive, D. and P. Voytek, Evidence for chemically-induced  structural
       gene mutations at the thymidine kinase locus in cultured  L5178Y
       mouse lymphoma cells, Mutation Res.,  44  (1977) 269-278.

26.    Clive, D., K.O. Johnson, J.F.S. Spector, A.G.  Batson and  M.M.M.
       Brown, Validation and characterization of the  L5178Y/TK+/~ mouse
       lymphoma mutagen assay system, Mutation Res.,  59 (1979) 61-108.

27.    Cole, J. and C.F. Arlett, Ethyl methanesulphonate mutagenesis with
       L5178Y mouse lymphoma cells: A comparison of ouabain,  thioguanine
       and excess thymidine  resistance, Mutation Res., 34  (.1976) 507-526.

28.    Cole, J. and C.R. Arlett, Methyl methanesulphonate mutagenesis in
       L5178Y mouse lymphoma cells, Mutation Res.,  50 U978)  111-120.

29.    Couch, D.B. and A.W. Hsie, Mutagenicity and  cytotoxicity  of
       congeners of two classes of nitroso compounds  in Chinese  hamster
       ovary cells, Mutation Res., 57 (1978) 209-216.

30.    Couch, D.B., N.L. Forbes and A.W. Hsie, Comparative mutagenicity of
       alkylsulfate and alkanesulfonate derivatives in Chinese hamster
       ovary cells, Mutation Res., 57 (1978) 217-224.

31.    Cox,  R.  and W.K. Masson, The isolation and preliminary
       characterisation of 6-thioguanine-resistant mutants of human diploid
       fibroblasts, Mutation Res., 36 (1976) 93-104.

32.    Cox,  R.  and W.K. Masson, X-ray-induced mutation to 6-thioguanine
       resistance in cultured human diploid  fibroblasts,  Mutation Res., 37
       (1976) 125-136.

33.    Davidson, R.L.,  K.A. O'Malley, and T.B. Wheeler,  Polyethylene
       Glycol-induced mammalian cell hybridization:  Effect of polyethylene
       glycol molecular weight and concentration, Somat.  Cell Genet.,  2
       (1976) 271-280.
                                      54

-------
34.    Dickerman, L.H. and J.A. Tischfield, Comparative effects of adenine
       analogs upon metabolic cooperation between Chinese hamster cells
       with different levels of adenine phosphoribosyltransferase activity,
       Mutation Res., 49 (1978) 83-94.

35.    Duncan, M.E. and P. Brookes, The induction of azaguanine-resistant
       mutants in cultured Chinese hamster cells by reactive derivatives of
       carcinogenic hydrocarbons, Mutation Res., 21 (1978) 107-118.

36.    Elsevier, S.M., R.S. Kucherlapati, G.A. Nichols, R.P. Creagan, R.G.
       Giles, F.H. Ruddle, K. Willecke, and J.K. McDougall, Assignment of
       the gene for galactokinase to human chromosome 17 and its regional
       localization to band q21-22, Nature, 251 (1974) 633-636.

37.    Fenwick, R.G. Jr. and C.T. Caskey, Mutant Chinese hamster cells with
       a thermosensitive hypoxanthine-guanine phosphoribosyltransferase,
       Cell, 5 (1975) 115-122.

38.    Fox, M., Factors affecting the quantitation of dose-response curves
       for mutation induction in Vyg Chinese hamster cells after exposure
       to chemical and physical mutagens, Mutation Res., 29 (1975) 449-466.

39.    Fox, M. and M. Radacic, Adaptational origin of some purine-analogue
       resistant phenotypes in cultured mammalian cells, Mutation Res., 49
       (1978) 275-296.

40.    Frantz, C.N., Bromodeoxyuridine resistance induced in mouse lymphoma
       cells by microsomal activation of dimethylnitrosamine, J. Toxicol
       Environ. Health 2 (1976) 179-187.

41.    Friedrich, U. and P. Coffino, Mutagenesis in S49 mouse lymphoma
       cells: Induction of resistance to ouabain, 6-thioguanine, and
       dibutyryl cyclic AMP, Proc. Natl. Acad. Sci. (U.S.), 74 (1977)
       679-683.

42.    Fujiwara, Y., Toshikazu, 0., and C. Heidelberger, Fluorinated
       Pyrimidines XXXVII, Effects of 5-Trifluoromethyl-2-deoxyuridine on
       the Synthesis of Deoxyribonucleic Acid of Mammalian Cells in
       Culture, Mol. Pharmacol., 6 (1970) 273-280.

43.    Heidelberger, C., Fluorinated Pyrimidines, Prog. Nucl. Acid Res.
       Mol. Biol., 4 (1965) 1-50.

44.    Hsie, A.W., P.A. Brimer, T.J. Mitchell and D.G. Gosslee, The
       dose-response relationship for ethyl methanesulfonate-induced
       mutations at the hypoxanthine-guanine phosphoribosyl transferase
       locus in Chinese hamster ovary cells, Somat. Cell Genet., 1 (1975;
       247-261.
                                      55

-------
45.    Hsie, A.W., P.A. Brimer,  T.J. Mitchell  and  D.G.  Gosslee,  The
       dose-response relationship  for  ultraviolet-light-induced  mutations
       at  the  hypoxanthine-guanine  phosphoribosyltransferase  locus  in
       Chinese hamster ovary  cells, Somat. Cell Genet.,  1  (1975)  383-389.

46.    Hsie, A.W., A.P. Li, and  R.  Machanoff,  A fluence  response  study  of
       lethality  and mutagenicity  of white, black,  and  blue fluorescent
       light,  sunlamp, and sunlight irradiation in  Chinese hamster ovary
       cells, Mutation Res.,  45  (1977) 333-342.

47.    Hsie, A.W., P.A. Brimer,  R.  Machanoff and M.H. Hsie, Further
       evidence for the genetic  origin of mutations in mammalian  somatic
       cells: The effects of  ploidy level and  selection  stringency on
       dose-dependent chemical mutagenesis to  purine analogue resistance in
       Chinese hamster ovary  cells, Mutation Res.,  45 (1977)  271-282.

48.    Huang, S.L., Dilution  of  hypoxanthine-guanine phosphoribosyl
       transferase and other  factors affecting the  frequency  of
       6-thioguanine resistance  in Chinese hamster  lung  cells, Mutation
       Res., 44 (1977) 119-128.

49.    Huberman, E. and L. Sachs, Mutability of different genetic loci  in
       mammalian cells by metabolically activated carcinogenic polycyclic
       hydrocarbons, Proc. Natl. Acad. Sci. (U.S.;, 73  (1976) 188-192.

50.    Huberman, E. and C. Heidelberger, The mutagenicity to mammalian
       cells of pyrimidine nucleoside analogs, Mutation Res., 14  (.1972)
       130-132.

51.    Jacobs, L. and R. DeMars, Quantification of  chemical mutagenesis in
       diploid human fibroblasts: Induction of azaguanine-resistant mutants
       by N-methyl-N1-nitro-N-nitrosoguanidine, Mutation Res., 53 (1978)
       29-53.

52.    Jones, G.E. and P.A. Sargent, Mutants of cultured Chinese hamster
       cells deficient in adenine phosphoribosyl transferase, Cell, 2
       (1974) 43-54.

53.    Jostes, R.F., L.E.  Hopwood, W.C. Dewey and G.R. Blackburn, The
       dependence of mutation frequency on 8-azaguanine concentration in
       control and irradiated CHO cells, Mutation Res.,  50 (1978) 433-440.

54.    Khalizev,  A.E.,  A.P. Pyzhov and N.I. Shapiro, Spontaneous
       mutagenesis in mammalian somatic cells and DNA replication.  II.
       Expression of gene mutations in cultured cells, Sov. Genet., 12
       (1976) 1345-1352.

55.    Kit, S., D.R. Dubbs, L.J. Piekarski, and T.C. Hsu, Deletion of
       thymidine kinase activity from L cells resistant  to
       bromodeoxyuridine,  Exp. Cell Res.,  31 (1963) 297-312.
                                      56

-------
56.    Kit, S., D.R. Dubbs, and P.M. Frearson, HeLa  cells  resistant  to
       bromodeoxyuridine and deficient in  thymidine  kinase  activity,  Int.
       J. Cancer, 1 (1966) 19-30.

57.    Kit, S., W.C. Leung, and D. Trkula, Properties of thymidine kinase
       enzymes isolated from mitochondrial and cytosol fractions of  normal,
       bromodeoxy-uridine-resistant, and virus-infected cells, in:   M. A.
       Mehlman & R. W. Hanson (Eds.) Control Processes in Neoplasia,
       Academic Press, New York, 1974, pp. 103-145.

58.    Knaap, A.G. and J.W.I.M. Simons, A mutational assay  system for
       L5178Y mouse lymphoma cells using hypoxanthine-guanine
       phosphoribosyl transferase (HGPRT)  deficiency as marker the
       occurrence of a long expression time for mutations  induced by x-rays
       and EMS, Mutation Res., 39 (1975) 97-110.

59.    Kozak, C., E. Nichols, and F.H. Ruddle, Gene  linkage analysis in  the
       mouse by somatic cell hybridization:  Assignment of  adenine
       phosphoribosyltransferase to chromosome 8 and a-galactosidase to  the
       X chromosome, Somat. Cell Genet., 1 (1975) 371-382.

60.    Kozak, C.D. and F.H. Ruddle, Assignment of the genes for thymidine
       kinase and galactokinase to Mus musculus chromosome  11 and the
       preferential segregation of this chromosome in Chinese hamster/mouse
       somatic cell hybrids, Somat. Cell Genet., 3 (1977)  121-133.

61.    Krahn, D.F. and C. Heidelberger, Liver homogenate-mediated
       mutagenesis in Chinese hamster V79 cells by polycyclic aromatic
       hydrocarbons and aflatoxins, Mutation Res., 46 (.1977) 27-44.

62.    Lea, D.E. and C.A. Coulson, The distribution of the numbers of
       mutants in bacterial populations, J. Genet., 49 (1948) 264-284.

63.    Lowry, O.K., N.J. Rosebrough, A.L. Farr and R.J. Randall, Protein
       measurement with the folin phenol reagent, J. Bioi. Chem., 193
       (1951) 265-275.

64.    Luria, S.E. and M. Delbruck, Mutations of bacteria from virus
       sensitivity to virus resistance, Genetics, 28 (1943) 491-511.

65.    Mayer, V.W. and W.G. Flamm, Legislative and technical aspects of
       mutagenicity testing, Mutation Res., 29 (1975) 205-300.

66.    Miller, O.J., P.W. Allderdice, D.A. Miller, W.R. Breg, and B.R.
       Migneon, Human thymidine kinase locus:  Assignment to chromosome 17
       in a hybrid of man and mouse cells, Science, 173 (1971) 244-245.

67.    Moore-Brown, M.M., D. Clive, B.E. Howard, A.G. Batson and K.O.
       Johnson, The utilization of trifluorothymidine (TFT) to select for
       thymidine kinase-deficient (TK"1"/") L5178Y mouse lymphoma cells,
       in preparation.
                                      57

-------
68.    Morrow J., Gene  inactivation  as  a mechanism  for  the  generation  of
       variability in somatic  cells  cultivated  in vitro, Mutation Res., 44
       (1977) 391-400.

69.    Myhr, B.C. and J.A. DiPaolo,  Requirement  for cell dispersion  prior
       to selection of  induced azaguanine-resistant colonies of Chinese
       hamster cells, Genetics, 80 (1975)  157-169.

70.    Newbold, R.F., P. Brookes, C.F.  Arlett,  B.A. Bridges and B. Dean,
       The effect of variable  serum  factors and  clonal morphology on the
       ability to detect hypoxanthine guanine phosphoribosyl transferase
       (HPRT) deficient variants in  cultured Chinese hamster cells,
       Mutation Res., 30 (1975) 143-148.

71.    Nikaido, 0. and M. Fox, The relative effectiveness of 6-thioguanine
       and 8-azaguanine in selecting resistant mutants from two V79  Chinese
       hamster cells in vitro, Mutation Res., 35 (1976) 279-288.

72.    O'Neill, J.P. and A.W. Hsie, Chemical mutagenesis of mammalian  cells
       can be quantified, Nature 269 (1977) 815-817.

73.    O'Neill, J.P. and A.W. Hsie, Phenotypic expression time of
       mutagen-induced 6-thioguanine resistance  in Chinese hamster ovary
       cells (CHO/HGPRT system), Mutation Res.,  59 (1979) 109-118.

74.    O'Neill, J.P., P.A. Brimer, R. Machanoff, G.P. Hirsch and A.W.  Hsie,
       A quantitative assay of mutation induction at the
       hypoxanthine-guanine phosphoribosyl transferase locus in Chinese
       hamster ovary cells (CHO/HGPRT system): Development and definition
       of the system, Mutation Res.,  45 U977) 91-101.

75.    O'Neill, J.P., D.B. Couch,  R.  Machanoff, J.R. San Sebastian,  P.A.
       Brimer and A.W.  Hsie,  A quantitative assay of mutation induction at
       the hypoxanthine-guanine phosphoribosyl transferase locus in Chinese
       hamster ovary cells (CHO/HGPRT system): Utilization with a variety
       of mutagenic agents, Mutation Res., 45 (1977) 103-109.

76.    O'Neill, J.P., J.C. Fuscoe  and A.W. Hsie, Mutagenicity of
       hetrocyclic nitrogen mustards  (ICR compounds) in cultured mammalian
       cells, Cancer Res., 38 (1978)  506-509.

77.    Orkin, S.H. and  J.W. Littlefield, Nitrosoguanidine mutagenesis in
       synchronized hamster cells, Exptl.  Cell Res., 66(1971)  69-74.

78.    Peterson,  A.R.,  D.F. Krahn, H. Peterson, C.  Heidelberger, B.K.
       Bhuyan and L.H.  Li, The influence of serum components on the growth
       and mutation of  Chinese hamster cells in medium containing
       8-azaguanine,  Mutation Res.,  36 (1976) 345-356.
                                      58

-------
79.    Reyes, P. and C. Heidelberger, Fluorinated pyrimidines, XXVI,
       Mammalian thymidylate synthetase:  Its mechanism of action and
       inhibition by fluorinated nucleotides, Mol. Pharmacol., 1 (1965)
       14-30.

80.    Roufa, D.J., B.N. Sadow and C.T. Caskey, Derivation of TK- clones
       from revertant TK+ mammalian cells, Genetics, 75 (1973) 515-530.

81.    Scheffe, H., Practical solutions of the Behrens-Fisher problem, J.
       Am. Stat. Assoc., 65 (1970) 1501-1508.

82.    Shapiro, N.I., A.E. Khalizev, E.V. Luss, E.S. Manuilova, O..N.
       Petrova and N.B. Varshaver, Mutagenesis in cultured mammalian
       cells.  II. Induction of gene mutations in Chinese hamster cells,
       Mutation Res., 16 (1972) 89-101.

83.    Shaw, C.I. and A.W. Hsie, Conditions necessary for quantifying ethyl
       methanesulfonate-induced mutations to purine-analogue resistance in
       Chinese hamster V79 cells, Mutation Res., 51 (1978) 237-254.

84.    Simons, J.W.I.M., Dose-response relationships for mutants in
       mammalian somatic cells in vitro, Mutation Res., 25 (1974) 219-227.

85.    Skopek, T.R., H.L. Leiber, B.W. Penman and W.G. Thilly, Isolation of
       a human lymphoblastoid line heterozygous at the thymidine kinase
       locus:  Possibility for a rapid human cell mutation assay, Biochem.
       Biophys. Res. Comm., 84 (1978) 411-416.

86.    Slack, C., R.H.M. Morgan, B. Carritt, P.S.G. Goldfarb, and M.L.
       Hooper, Isolation and characterisation of Chinese hamster cells
       resistant to 5-fluorodeoxyuridine, Exper. Cell Res., 98 (1976) 1-14.

87.    Taylor, M.W., J.H. Pipkorn, M.K. Tokito, and R.O. Pozzatti Jr.,
       Purine mutants of mammalian cells III.  Control of purine
       biosynthesis in adenina phosphoribosyl transferase mutants of CHO
       cells, Somat. Cell Genet., 3 (1977) 195-206.

88.    Thacker, J., M.A. Stephens and A. Stretch, Factors affecting the
       efficiency of purine analogues as selective agents for mutants of
       mammalian cells induced by ionising radiation, Mutation Res., 35
       (1976) 465-478.

89.    Thacker, J., A. Stretch and M.A. Stephens, The induction of
       thioguanine-resistant mutants of Chinese hamster cells by  -rays,
       Mutation Res., 42 (1977) 313-326.

90.    Thacker, J., M.A. Stephens, and A. Stretch, Mutation to
       ouabain-resistance in Chinese hamster cells:  Induction by ethyl
       methanesulphonate and lack of induction by ionising radiation,
       Mutation Res., 51 (1978) 255-270.
                                      59

-------
91.    Thilly, W.G., J.G. Deluca, H. Hoppe  IV and B.W. Penman,  Phenotypic
       lag and mutation  to 6-thioguanine resistance  in diploid  human
       lymphoblasts, Mutation Res., 50  (1978) 137-144.

92.    Thompson, L.H., S. Fong, and K.  Brookman, Validation of  conditions
       for efficient detection of HPRT  and APRT mutations in suspension
       cultured Chinese  hamster ovary cells, Mutation Res., in  press.

93.    Tischfield, J.A.  and F.H. Ruddle, Assignment  of the gene tor adenine
       phosphoribosyltransferase to human chromosome 16 by mouse-human
       somatic cell hybridization, Proc. Natl. Acad. Sci. USA,  71  (1974)
       45-49.

94.    Umeda, M. and C.  Heidelberger, Comparative studies of fluorinated
       pyrimidines with  various cell lines, Cancer Res., 28 (1978).

95.    van Zeeland, A.A. and J.W.I.M. Simons, The effect of calf serum on
       the toxicity of 8-azaguanine, Mutation Res., 27 (1975) 135-138.

96.    van Zeeland, A.A. and J.W.I.M. Simons, Linear dose-response
       relationships after prolonged expression times in V-79 Chinese
       hamster cells, Mutation Res., 35 (1976) 129-138.

97.    van Zeeland, A.A. and J.W.I.M. Simons, The use of correction factors
       in the determination of mutant frequencies in populations of human
       diploid skin fibroblasts, Mutation Res. 34 (1976) 149-158.

98.    Westerveld, A., R.P. Visser, P. Meeva Khan,  and D. Bootsma, Loss of
       Human genetic markers in man-Chinese hamster somatic cell hybrids,
       Nature New Biol., 234 (1971) 20-24.
                                     60

-------
AAr

AGr

aprt+/+

aprt+/~


aprt~/~

C HO- IB- 2

CHO-AT3-2

hgprt"1"


hgprt"

OUAR

FUdRr

PE

S~,  cells
WT
                     GLOSSARY

cell phenotype, resistance  to AA

cell phenotype, resistance  to AG

cell genotype, with  two  genes for  functional APRT enzyme

cells with only one  gene  (heterzygous)  tor functional APRT
enzyme

cells with no gene for functional  APRT  enzyme

cell line heterzygous at  the  aprt  locus

cell line heterzygous at  the  aprt  and ^k loci

cell genotype, with  a gene  (X-linked) for functional HGPRT
enzyme

cells without a gene for  functional  HGPRT enzyme

cell phenotype, resistance  to QUA

cell phenotype, resistance  to FUdR

ratio of viable ells to cells plated; mean plating efficiency

surviving cell fraction; mean cell survival relative to
controls

cell phenotype, resistance  to TG

cell genotype, with  two genes for  functional TK  enzyme

cell genotype, with  only one  gene  for functional TK enzyme

cells with no gene for functional  TK enzyme

wild type cells (aprt"1"/"1", hgprt"1",  tk+/+,  ATPase  with
normal sensitivity to ouabain)
                                      61

-------
                                   TECHNICAL REPORT DATA
                            (f lease rf.acl Instructions on t/ic reverse before cutnplctinx}
1. REPORT NO.
                              2.
4. T.TLE AND SUBTITLE   Mutagenicity Testin- in Mammalian
 Cells.  The development and Validati.o'.i of Multiple
 Drug-Resistance Markers Having Practical Application
 for Screening  Potential Mutagens.	
                                                           3. RECIPIENT'S ACCESSION NO.
             0. REPORT DATE
             6. PERFORMING ORGANIZATION CODE
7. AUTHOR(S)
                                                           8. PERFORMING ORGANIZATION REPORT NO
 June II. Carver, Gerald  M.  Adair, and Daniel L. Wandres
9. PERFORMING ORGANIZATION NAME AND ADDRESS

 Lawrence Livermore  Laboratory
 Biomedical Sciences Division
 University of California
 Livermore. CA  94550	
             10. PROGRAM ELEMENT NO.

               1NE 625F
             11. CONTRACT/GRANT NO.
               EPA-IAG-D5-E681-AN and  AO
12. SPONSORING AGENCY NAME AND ADDRESS
U.  S.  Environmental Protection  Agency
Washington,  D.C.  20460
             13. TYPE OF REPORT AND PERIOD COVERED
               Interim 1/76 thru  1/79
             14. SPONSORING AGENCY CODE

               EPA/600/17
15. SUPPLEMENTARY NOTES
         This project  is part of the EPA-planned  and  coordinated Federal  Interagency
         Energy/Environment R&D Program.
16. ABSTRACT
       Chinese hamster ovary  (CHO)  cell lines heterozygous at  the  adenine phospho-
 ribosyltransferase  (aprt)  and  thymidine kinase (tk) loci were  isolated and used for
 single-step selection of spontaneous and induced mutants resistant  to 8-azaadenine
 (AAr), 6-thioguanine (TGr),  5-fluorodeoxyuridine (FUdRr), or  ouabain (OUAR).   Mutation
 data are reported for direct mutagens ethyl methanesulfonate  (EMS),  N-methyl-N'-nitro-
 N-nitrosoguanidine  (MNNG), 4-nitroquinoline-l-oxide (NQO) and  for promutagens dimethyl-
 nitrosamine (DMN) and benzo(a)pyrene (BaP) activated by rat liver homogenates.
 Critical plating densities were  established for AAr, TGr, and  FUdRr.   Optimal ex-
 pression of mutant  phenotypes  after mutation induction with EMS, DMN,  or BaP were 2 to
 4 d for AAr, 6 to 8 d for  TGr,  3 d for OUAR,  and 1 to 3 d for  FUdRr.   The induced
 mutant frequencies  as a function of relative cell survival after  treatment with EMS,
 DMN, or BaP showed  locus-specific  differences in sensitivity.  Of 61  clonal isolates
 resistant to AA and assayed  for  APRT activity, 87% had < 5% wild type  activity; of 30
 TGr clones assayed, 83% had<5%  wild type HGPRT activity.  Of  42  FUdRr clones assayed,
 41 had
-------