FURTHER STUDIES OF THE SUPPRESSOR-MUTATOR SYSTEM OF
        CONTROL OF GENE ACTION IN MAIZE
             Barbara McCZintock
For many years, evidence of the presence  studies of maize, and these components
in the chromosome complement of dis-  were designated "controlling elements."
tinct genetic components that control the  Recently, however, controlling elements
action of genes was available only from  have been identified in bacteria, where it


470   CARNEGIE INSTITUTION OF WASHINGTON

is possible to explore at the chemical level
some aspects of their mode of action. It is
expected that continued examination of
these elements in bacteria will lead to an
appreciation of the mechanism of their
operation, with a precision that would be
difficult or impossible to duplicate in maize.
Nevertheless, at the phenotypic level, the
performances of the bacterial and maize
control systems exhibit parallels that sug-
gest basic similarities in gene control in
these two widely separated organisms.
The bacterial systems, as described by Jacob
and Monod, are composed of two genetic
elements, each distinct from the "struc-
turalfl gene. One of them, termed the
"operator," is adjacent to the structural
gene (or sequence of structural genes),
and directly controls its activation. The
structural gene is considered to carry the
code that is responsible for a particular
sequence of amino acids and thus for the
specificity of a protein. When the struc-
tural gene is activated, this protein is
formed. The second element of the system,
termed the "regulator," may be located
either near the structural gene or elsewhere
in the bacterial chromosome. The regula-
tor is responsible for the production of a
repressor substance, not a protein, that ap-
pears in the cytoplasm. The operator ele-
ment responds in some yet unknown man-
ner to a change in degree of effective action
of the repressor substance by turning on or
turning off the functioning of the struc-
tural gene in accordance with such change.
Each operator-regulator system is specific,
in that an operator will respond only to the
particular product of the regulator of its
system.

In maize, likewise, some of the control
systems are composed, basically, of two
elements. One, closely associated with the
structural gene and directly controlling its
action, can be likened to the operator ele-
ment in bacteria. The other, which may
be located near the first element or may be
independently located in the chromosome
comolement. establishes the conditions to

which the gene-associated element  rc-
sponds, a particular change in these con.
ditions being reflected in a particular
change in action of the gene. It thus is
comparable to the regulator element in
bacteria. In maize, as in bacteria, each
operator-regulator system is quite specific:
an operator element will respond only to
the regulator element of its own system.
Discovery of the presence of controlling
elements in maize was made possible bv
the fact that under certain conditions 3
controlling element may be transposed
from one location to another in the chro-
mosome complement. Transposition of the
operator-like element to the locus of a gent
will bring the action of that gene under
the control of the system to which the
element belongs.

During the year, studies of one of thcst
systems were extended. The element th;lt
is comparable to the regulator in this
system is designated Suppressor-mutator
(Spm), and the system has been refer&
to in the past as the Spm system of con-
trol of gene action.

Knowledge of the function of the Splt~
system was initially obtained by studying
gene action at the A1 locus in chromosome
3 and the A, locus in chromosome 5 after
the operator element of the system h:l(l
been inserted at each of these loci. The
modified loci were designated al"`-' and
a2 m-1. Three additional inceptions of con-
trol of gene action by this system have no\\'
been investigated. TWO of them ag:lill
involve the standard A1 locus in chroml-
some 3, and are designated al"'-* and al"`-
The symbols m-l, m-2, and m-5 refer [I'
the order in time of inception of contr(ll
of gene action of A1 by the Spm s\`s[c~ll
(m-3 and m4 refer to inceptions oi toll
trol of gene action at A1 by the Dj-Ac col'-
trol system). The third inception involvt"
the locus of Wx in chromosome 9, and thy
modified locus is designated ZUX"+~. (Bori'
A, and A, are associated with anthocYi:'
pigment formation in plant and kerntl.
Wx is associated with oroduction of am! `-


DEPARTMENT OF GENETICS   471

Ljc in the pollen grain and the endosperm  null expression, but most of them produce
j the kernel.) The type of evidence that  pigment in the aleurone layer; the intensity
>.l~ allowed recognition of control of gene  of pigmentation, ranging from faint to
,!ction at al"+*, als5, and ur.P8 by. the  rather deep, distinguishes the mutants from
jp" system will be reviewed in the fol-  one another. All type-2 mutants that pro-
IJwing sections.            duce pigment, however, are characterized
                 by differences in intensity among indi-

Control of al-' by the Spm System   vidual cells of the aleurone layer: the layer

Early in the study of at*`, an ear of a
$ant that was Al/al (standard recessive);
,j2/az in constitution was utilized in a
;toss with a plant that was homozygous
ior the standard recessive, al, and for the
standard dominant, A*. (The standard
recessive, al, is completely stable in the
ljresence of Spm. There is an operator
component at the locus of standard al, but
It belongs to the control system of which
Dotted, Dt, is the regulator. The a2 in the
heterozygous parent plant was derived
trom mutation at az"+l, which had oc-
curred in a cell whose nucleus also carried
nctive Spm.) The phenotypes of the
kernels on the ear produced by this cross
were those expected, with one exception.
Instead of being either fully pigmented or
totally colorless, the exceptional kernel ex-
hibited spots of pigment in a nonpig-
mented background, suggesting that the
A1 locus had been modified in a cell of
the ear-bearing parent plant, whose prog-
eny cells gave rise to the kernel. The
plant derived from this kernel also ex-
hibited variegation for anthocyanin pig-
mentation, and tests conducted with it con-
firmed the presence of a modified A1 locus,
which was then designated al"`-`.
Initial tests of al"`, conducted some
years ago, showed that mutations occur
at this locus and that they result in two
distinct types of mutants, whose subse-
quent expressions are stable. Gene action,
in both plant and kernel, expressed by one
type (type 1) resembles that produced by
the standard Al. Mutants of the second
type (type 2) occur far more frequently,
and fall into a graded series with respect
to ability to form anthocyanin pigment.
A small percentage of them exhibit the

is speckled with cells in which the pigmen-
tation is much more intense than in sur-
rounding cells. The expression of type-l
and type-2 mutants also differs in the plant.
With respect to distribution of anthocyanin
pigment to various parts of the plant, and
response of the pigment-producing system
to sunlight, the expression of type-l mu-
tants resembles that associated with stand-
ard Al. When a type-2 mutant that
produces pigment is present, the pigment is
restricted to certain parts of the plant, and
the response of the pigment-producing
system to sunlight is much retarded.
The early studies of al"-' were con-
ducted with only three successive genera-
tions of plants, and the system responsible
for control of its behavior was not then
identified. Some years later when the Spm
system  had been identified through
analyses of aIs1 and a2-l behavior, it was
suspected that this system was also re-
sponsible for control of gene action at
al**. Consequently, examination of al**
was recommenced, and a number of ex-
ploratory tests were conducted with it dur-
ing the past year. They revealed that gene
action at al"+' is under the control of the
Spm system. The results also suggest that
in this instance Spm is located close to the
structural gene(s) of the Al locus, and
that its transposition away from that locus
is associated with the induction of the
above-described mutants. It is possible
here, as with b.P' (see Year Book 55), that
the operator and regulator components
both reside at or near the locus of the
structural gene, and that mutation-induc-
ing events are usually associated with re-
moval of both elements from the vicinity of
the locus. If so, removal of the regulator


472   CARNEGIE INSTITUTION OF WASHINGTON

element Spm without concomitant re-
moval of the operator element should occur
on rare occasions, and a search that may
reveal this event is under way.

The conclusion that the Spm system
controls gene action at al'+' is derived
from several types of observation. In the
first place, Sprn was present in each of 44
a1 *`-carrying plants that were tested for
its presence. Its location was very close to
al** in plants that had only one Spm
element; in plants having two such ele-
ments, one was located close to al-*. Some
plants had three or more independently
located Spm elements, but when so many
were present the tests did not give con-
clusive evidence of their relative locations.
The second pertinent observation concerns
Spm constitution and location in plants
carrying a stable mutant of al**. In
crosses of al*`-carrying plants to plants
homozygous for standard al and having no
Spm, some kernels carrying a germinal
mutation may be present on the ears in
addition to those that have received un-
modified al*`. Tests of the presence or
absence of Spm in plants from both types
of kernels revealed that Spm was present
in each plant derived from a kernel carry-
ing an unmodified al"+*, and that when
only one Spm was present it was located
close to al*`. In contrast, some of the
plants derived from the mutant class of
kernels had no Spm, whereas in others,
although Spm was present, it was never
found to occupy a position adjacent to the
10~~s of the mutant.

The third pertinent observation is con-
cerned with Spm constitution in the parts
of ears of al-`/al plants that are derived
from cells in which a mutation at al"+*,
producing a stable allele, has occurred early
in plant development. If such an ear is
tested for the presence or absence of Spm
in individual kernels, it is possible to learn
whether this element was present in the
cells that produced the mutant sector, and,
if so, its position in relation to the mutant
locus. On some ears exhibiting such
mutant sectors, no evidence' of Spm was

observed among the kernels within the
sector, although it was present in kern&
on other parts of the same ear. On other
ears, spm was present in the cells that
produced the sector but was not close]!
linked with the mutant phenotype.
The above-described observations con.
formed with the assumption that Sptn i$
responsible for control of gene action at
aI** ; but the decisive evidence was oh-
tained from observations of the response
of al"" to change in phase of activity o[
Spm. When Spm is in its inactive phase,
no gene action at al'+* is expressed in either
plant or kernel, and consequently antho-
cyanin pigment is absent. When Spnr is in
its active phase, however, mutations occur
that permit production of anthocyanin.
Moreover, if a plant or kernel begins tic-
velopment with Spm in an inactive phase,
the time of occurrence of such mutation
at a," is a function of the time during
development when Spm changes to an ac-
tive phase: the later the time of chan:c,
the later the occurrence of mutation.

A Third Znception of Control of Gent

Action at the A1 Locus by the
     Spm System
In the course of investigation of the

Ds-AC system of control of gene action, il
was necessary to maintain a stock cultLlr<
carrying a particular combination of gellc
markers and also AC. This stock \VJ\
homozygous for the standard domin:ll~!
alleles of all but one (i.e., Bz) of the SC[`C
loci associated with anthocyanin pipmell[
formation. To maintain this stock, Silj
crosses were made each year. In 1957, 21'
ear produced by one of these sib crosses rC-
vealed the presence among some of 1:'
kernels of somatically occurring chase 1"
action of a gene concerned with anthCl-
cyanin pigment formation. The folloivin.z
year, testcrosses were made with a is"
plants grown from these kernels, to dctc'-
mine whether the locus involved was On"
that had previously . been identified, I'
proved to be A1 in chromosome 3; s!iL'


DEPARTMENT OF GENETICS   473

:he modified locus was designated ale6.
Other crosses with these plants indicated
:hat the AC system was not in control of
<ene action at al'))-`. Spm, however, was
iound to be present in each aim-5-carrying
[`Iant that exhibited somatically occurring
;banges in gene action; and the ratio of
kernel types on ears produced by test-
;rosses with these plants suggested that
gene action at al*' might be under the
control of the Spm system. Tests con-
ducted during the past year have shown
this to be so.
The originally isolated state of alG5, in
the presence of an independently located,
Active Spm, is characterized by the pro-
duction of somatic mutations to stable
alleles of Al, often occurring early in
plant or kernel development. Most of them
fall into two distinct groups: those that
express high levels of AI gene action, and
those that express either low levels or the
null level. Mutants of these two main
types occur with nearly equal frequencies,
and both are stable in the presence of ac-
tive Spm. The original state of al"+' also
gives rise to new states in the presence of
active Spm, either class I or class II.
The original state of al"+" produces so
many early-occurring mutations in the
presence of active Spm that it was not as
useful as some of the derived states for
the purpose of determining whether or
not the Spm system is responsible for con-
trol of gene action at al"+`. The derived
class I states that give rise to mutations
only late in development of plant and
kernel served this purpose well, however,
and four such states were selected for tests,
It was learned that when Spm is absent (or
present in an inactive phase) each of these
states is characterized by a low level of
A, gene action, in both plant and kernel,
an expression that continues as long as
Spm is absent or inactive; but in the
presence of active Spm all gene action is
suppressed, until in some cells, late in the
development of plant or kernel, a mutation-
inducing event at cl,"`-' allows some level of
A1 gene action to be expressed.

A class II state of al"+' was also ex-
amined. In the absence of active Spm, this
state produces a low level of gene action
in both plant and kernel, whereas in the
presence of active Spm all gene action is
suppressed. With this state, however, in
contrast to the class I states, no mutation-
inducing events or alterations of state
were observed. It behaves as a typical class
II state.

One unusual state of al"+' was identified
and further examined. In the absence of
active Spm, this state expresses a low level
of A1 gene action in both plant and kernel.
When an active Spm is present, all gene
action is suppressed but alterations of al'"-'
do occur, many of them resulting in the
appearance of new states. The new states
are distinguished from one another by the
types of mutation they give in the presence
of active Spm, by the relative frequency of
occurrence of each type (if more than one
type is produced), by the time during de-
velopment when the mutations occur, and
by the frequency of occurrence of mutation
at any one time during development. Thus
an array of new states has been derived
from this one state.

Control of Gene Action at the Locus of

Wx by the Spm System

During examination of the relative fre-
quencies of production of different types
of stable mutants of A2 given by the
original state of a2-l in the presence of
active Spm, an ear of a plant having the
constitution a2"+l bt/aa Bt; Wx/wx was
used in a cross with a plant that was
homozygous for a?, bt, and wx and had
no Spm. (a2 = standard recessive, stable
in the presence of active Spm; bt = brittle
endosperm, located approximately 6 cross-
over units from the locus of A2 or its
alleles.) The kernel types on the resulting
ears were those expected, with one ex-
ception. The exceptional kernel was color-
less and Bt, but was neither totally WX
nor totally wx. The endosperm was a
mosaic of sectors exhibiting different levels


of Wx gene act&, from the null level
  (WX) in some sectors to the full Wx level
  in others, the pattern resembling that pro-
  duced by the particular state of a2-l in the
  ear-bearing parent plant in the presence
  of active Spm. It was suspected that the
  operator element of the Spm system had
  been incorporated at the Wx locus in
  chromosome 9 in a cell of this plant, and
  that the descendants of this cell, which
  gave rise to the exceptional kernel, ex-
  pressed control of gene action at the Wx
  locus'by the Spm system. A plant was
  g&wn from the exceptional kernel (plant
  `number 7774) and subjected to tests that
/  would determine whether or not the hy-
/   pothesis was correct. The results gave clear
  evidence of control by the Spm system of
  gene action at wx+`+*, as the newly modi-
  fied Wx locus was designated.
   Plant 7774 proved to be a2 &/a2 bt,
   wx- /wx in constitution, and it had two
  independently located active Spm elements
  in most of its tested parts. The plant pro-
  duced three tillers; and testcrosses were
  made with pollen from the main stalk and
  from each of the tillers. In addition, it
  produced four fertile ears, and a testcross
  was made with each. The types of test-
  cross that established Spm control of gene
   action at wxks were as follows:
   1. Reciprocal crosses with plants having
  no Spm that were homozygous for the
   standard, stable recessive wx and carried
  a class I state of a2"+l. Three distinctly
   different class I states of az"-l were utilized
    in this test.

2. Crosses to plants homozygous for wx
and carrying a class I state of a2-l and
also Spmw (a weak-acting Spm; see Year
Book 56).

3. Crosses to plants homozygous for wx
and carrying a class II state of a2*l. Some
of the plants had no Spm; others carried
Spm in an inactive phase of long duration.
In the kernels on the ears produced by
these crosses, aZ"+l served as an indicator
of the presence or absence of Spm, and of
its type of action if present. In each kernel

their behavior could be compared. Corm
respondence in response of azwl ancl
wx-* in these kernels was so complerc
that there could be no doubt of control of
wx"+* by the Spm system. They responded
in like manner to presence or absence of
active Spm, to Spmw, and to an inactive
Spnz that underwent change to the active
phase in a few cells late in kernel &.
velopment.

The behavior of this original isolate 01
-* may be summarized as follows: 111
:b
e a  sence of active Spm, it exhibits ;I
very low level of gene action, an expression
that remains stable as long as Spm is aI)-
sent or is present in an inactive phase.
With a fully active Spm, all gene action i$
suppressed until a mutation-inducing evcn1
at wx** results in a stable mutant. Sony
such mutants give rise to the apparent ft11l
Wx phenotype; others produce an intcr-
mediate level; and still others give a vcr\;
low level or the null expression. Such mu-
tations may occur in a cell either early OI
late in plant or kernel development. Plnn~
7774 exhibited some early-occurring mut+
tions. One of its ears, for example, w25
derived from a cell in which a mutntioll
had restored full Wx gene expression; nn(l
one of its tassels had a large sector carryin:
a stable mutant that gave the null (I~~.~ )
expression.

Control of Reversals in Spm Actiru't\
         Phase

An important aspect of the Spm comrl
system is the cyclically occurring change iI1
phase of activity of Spm-from active to
inactive and back to active. It was e.lrl\
noted that the duration of a phase is toll-
trolled in some manner. It may be shnrl.
lasting only a few cell generations, or i:
may extend through many cell gellcr.t-
tions or even plant generations. TtlL
early studies also suggested that cnn-
trol of the duration of a phase resi(lci.
primarily, in the event that initiates 11:~
reversal of phase. To obtain more evidcIl;c.

that had received both azm-l and wPs,  additional experiments were pertorIllcl:

474   CARNEGIE, INSTITUTION OF WASHINGTON


DEPARTMENT OF GENETICS   475

-

this year, all of them conducted with
plants descended from a single plant hav-
ing one Spm closely linked with the WX
;ene in chromosome 9. The behavior of
this particular Spm element has been ob-
served through six generations of plants,
nnd the initial studies of reversal of phase
of activity of Spm were conducted with it.
The studies this year had two objectives:
(1) to compare the times and frequencies
of occurrence of phase reversal of Spm
in sister plants, all carrying an Spm that
had undergone reversal of phase in a cell
of the mother plant; and (2) to determine
the stability of phases of Spm activity when
two elements were present in a plant,
occupying allelic positions but in alternate
phases of activity at the beginning of plant
development. In all tests, a class II state
of lr2"-1 was employed to detect the times
and frequency of occurrence of changes in
Spm activity phase, as described in previous
Year Books.

1. Kernels were selected from an ear
of each of five different plants, either
Wx +/wx Spm or Wx Spm/wx + in
constitution, in which a part of the ear had
developed from a cell in which a reversal
in phase of Spm activity had occurred.
Plants were grown from kernels within
the sector exhibiting the altered phase of
activity, and also from kernels that had
appeared on the remaining part of each
ear, where no such alteration had occurred.
Comparisons were made with regard to
initial phase of activity of Spm, as well as
times and frequency of occurrence of re-
versals of phase.  These observations re-
vealed marked similarities in pattern of
phase reversal among the plants derived
from kernels within a sector, contrasting
with the pattern in plants derived from
kernels on other parts of the same ear.
2. Plants were constructed that carried
one inactive Spm element, which for four
generations had exhibited a long duration
of the inactive phase, and in addition an
active Spm that was known to undergo
phase reversal late in plant and kernel
development. (Both elements, as was stated

earlier, were derived from a single ancestor
plant having one Spm closely linked to
Wx in chromosome 9.) In the tested
plants, the two elements occupied allelic
positions in chromosome 9, the inactive
Spm closely linked to Wx and the active
Spm closely linked to wx. TO test whether
or not, under these conditions, control of
phase reversal would continue to reside in
each Spm element itself, the ears produced
by each plant were crossed by plants that
were homozygous for wx and had no Spm,
and also by plants that were homozygous
for wx but carried Spm in its active phase.
The second cross reveals the presence of
inactive Spm, as described in Year Book 57.
The Wx and wx classes of kernels on the
resulting ears were compared with respect
to Spm activity phase. It was apparent that
control of the duration of activity phase
had not been modified in either Spm ele-
ment by their association at allelic positions
in the nuclei of the ear-bearing plants:
the characteristic duration of phase of each
Spm was exhibited in the progeny. This
evidence confirmed the previously drawn
conclusion that control of reversal of Spm
activity phase resides in the Spm element
itself.

Nonrandom Selection of Genes Coming
under the Control of the Spm System
Five independent inceptions of control
of gene action by the Spm system have
been examined so far: three at the locus
of Al in chromosome 3, one at the locus of
A, in chromosome 5, and one at the locus
of Wx in chromosome 9. A1 and Az are
associated with anthocyanin pigment for-
mation in plant and kernel, and Wx is
concerned with development of amyIose
starch in pollen and endosperm. There is
no reason to believe that control of gene
action by this system is confined only to
genes associated with anthocyanin produc-
tion or with the structure of starch. The
ease of detection of any change in action
of these genes is believed to be responsible
for the ready determination of the con-


476   CARNEGIE INSTITUTION OF WASHINGTON

trol system responsible for the change. It
should be recalled that early in the study
of controlling elements many independent
inceptions of control of action of genes as-
sociated with chlorophyll development
were recognized. Study of some of these
was initiated, but it was soon decided to
discontinue their examination in order to
concentrate on systems that affect genes as-
sociated with anthocyanin development in
the kernel or in both plant and kernel.
It was realized that much more could be
learned about the action of a controlling
element on a gene if the phenotypic ex-
pression of the gene was exhibited in both
plant and kernels or even in kernels alone.
The kernels are especially valuable in this
regard, because the phenotype of the endo-
sperm registers the genetic constitution in
the succeeding generation. Very large
numbers of kernels may be obtained in
progeny tests of a single individual, and
with little technical difficulty. To obtain
similar numbers of progeny plants would
introduce technical difliculties of some
magnitude, which might reduce the effi-
ciency of the studies. Before the studies
of control systems associated with chloro-
phyll genes were discontinued, however,

the system associated with one such gene
had been examined in some detail. The
locus concerned was named "mutable
luteus" (Itr"). As was mentioned in Year
Book 57, the system responsible for con-
trol of Iu"' resembles in detail the Spnr sys-
tern. Moreover, az*' originated in a plant
carrying 2~"`. It is very probable that /zim
represents a case of control by the Spr,l
system of a gene associated with chloro-
phyll production, although direct proof
has not been obtained.

In analyses of the mode of operation of
a gene-control system, the phenotypic es-
pression of the gene controlled is of parn-
mount importance, particularly during
initial examinations of the system. Once a
control system has been identified, through
its action on appropriate genes, and
methods have been devised that will
readily detect its action at other gene loci,
it becomes practicable to examine its con-
trol of genes whose phenotypic expres-
sions are less well adapted for such studies.
In the studies reported so far, genes that
are most appropriate for initial examinn-
tion of control systems have consisten+
been chosen.