THE
AMERICAN NATURALIST

 

Vol. XCV September-October, 1961 No. 884

SOME PARALLELS BETWEEN GENE CONTROL SYSTEMS
IN MAIZE AND IN BACTERIA

BARBARA MCCLINTOCK

Department of Genetics, Carnegie Institution of Washington,
Cold Spring Harbor, New York

It has been realized for some time that, although the gene is necessary
for expression of a certain phenotype, it may not in itself be sufficient for
such expression and mechanisms may exist that control its action. Genetic
systems that serve this purpose in maize were recegnized some years ago,
and studies conducted with a number of them have been reported (for refer-
ences, see Brink, 1958, 1960; McClintock, 1956a and b; Peterson, 1960).
Without adequate confirmation of similar systems in other organisms, it
could be considered that the systems in maize may not reflect a type of con-
trol of gene action that is common to organisms in general. Recently, how-
ever, genetic systems that control gene action have been discovered in
bacteria (Jacob and Monod, 1959, 1961; Jacob et al., 1960) and it is now
apparent that a relationship may exist between the bacterial and che maize
control systems. The bacterial control systems, described by Jacob et al.,
are composed of two genetic elements, each distinct from the ‘‘structural’’
gene. One of them, designated the ‘‘operator,’’ is located adjacent to the
structural gene (or sequence of structural genes) and controls its activation.
The structural gene, when activated, is responsible for the production of a
particular sequence of amino acids and thus for the specificity of a protein.
The second element of this system, termed the ‘‘regulator,’’ may be located
close to the structural gene, or it may be located elsewhere in the bacterial
chromosome. The regulator is responsible for the production of a repressor
substance—not a protein—that appears in the cytoplasm. The operator ele-
ment responds in some yet unknown manner tc changes in degree of effec-
tive action of the repressor substance by ‘‘turning on’’ or “‘eurning off’? the
action of the structural gene in accordance with such changes. Each
operator-regulator system is specific, in that an operator will respond only
to the specific product of the regulator of its system.

In maize likewise, some of the control systems are composed, basically,
of two elements. One is closely associated with the structural gene and di-
rectly controls its action; it may be likened to the operator element in bac-
teria. The other element may be located near the first or may be independ-
ently located in the chromosome complement. It establishes the conditions

265
266 THE AMERICAN NATURALIST

to which the gene-associated element responds, a particular change in these
conditions being reflected in a particular change in action of the gene, and
thus is comparable to the regulator element in bacteria. In maize, as in bac-
teria, each ‘‘operator-regulator’’ system is quite specific: an “‘operator’’
element will respond only to the particular ‘‘regulator’’ element of its own
system,

Several different two-element control systems, each operating independ-
ently of the others, have been identified in maize. These were discovered,
originally, because the elements belonging to each were transposed from
one location to another in the chromosome complement without losing their
individual identities in the process. It was found that the gene-associated
element of a system can leave the locus of one gene and become associated
with that of another. After such an association is established, the action of
the gene comes under the control of the system to which the gene-associated
element belongs. It has been possible, therefore, to examine the mode of
operation of a particular control system at a number of different gene loci
and, conversely, to examine the operation of different control systems at the
same gene locus. It should be emphasized that, although transposition of
controlling elements in maize made it possible to recognize their presence
in the chromosome complement and to study the mode of operation of the
component elements of a system, transposition does not necessarily charac-
terize the behavior of a controlling element. An element previously ex-
hibiting transposition may become fixed in location. If it is the gene-
associated element that becomes fixed, the action of the gene will then be

_ permanently under the control of the system to which that elem
Examples will be considered in this report.

Jacob (Jacob, 1960; Jacob, Schaeffer and Wollman, 1960) and Richter
(1961) have postulated that controlling elements in maize may be compara-
ble to episomes in bacteria, Recent evidence (Buttin, Jacob and Monod,
1960; Yarmoliasky and Wiesmeyer, 1960) about the manner in which a phage
particle may control the action of bacterial genes in the neighborhood of its

‘attachment to the bacterial chromosome lends support to this interpretation.
In a lysogenic bacterium, induction of phage by ultraviolet light or by chemi-
cal treatment releases inhibition of gene action not only in phage genes that
are concerned with initiating vegetative replication but also in genes of the
bacterial chromosome in the neighborhood of phage attachment. This effect
resembles that which occurs in maize when a controlling element at the
locus of a gene is removed by the transposition mechanism.
action of the gene accompanies this removal.

Notwithstanding the analogies that may be drawn between controlling ele-
ments in maize and episomes in bacteria, it now appears to the author that
control systems in maize also resemble the opetator-regulator systems of
control of gene action in bacteria, as outlined above. In maize, as in bac-
teria, the controlling element (the “‘operator’’) at the locus of the structural
gene responds to altered activities of the second element (the ‘
of the system by inducing modification in action of the structur

ent belongs.

A change in

‘regulator’’)
al gene. In
GENE CONTROL IN MAIZE AND BACTERIA 267

maize, the response of the ‘‘operator’’ element to change in effective action
of the ‘‘regulator’’ element results in controlled types of change in action of
the structural gene, and many such changes are accompanied by removal of
the ‘‘operator” element from the locus. In other cases, however, the ‘‘oper-
ator’’ element is not removed from the locus. It responds merely by ‘‘turning
on’’ or “turning off’? the action of the structural gene. When this occurs, the
maize systems resemble the operator-regulator systems in bacteria.

The control system composed of the elements Dissociation (Ds) and Acti-
vator (Ac) was the first of those in maize to be explored extensively. Its
mode of operation was examined at a number of different gene loci (McClin-
tock, 1953). This system was studied intensively because it was possible
to identify readily both the ‘operator’? element, Ds, and the “‘regulator’’
element, Ac. In several cases (bz™-4, sh™, c™—1) the proximal or distal
position of Ds with respect to the components of the structural gene could
also be determined. With some other two element systems, the regulator
element is readily identifiable but the presence of the operator element at
the gene locus often must be assumed on the basis of the control that the
regulator exerts on gene action. However, when adequate test methods are
available, it is possible to confirm the presence of an operator element at
the locus of the gene by means of crossover techniques which are capable
of defining its location with respect to the components of the structural
gene. This method has been used successfully to define the location of this
element in the case where action of the structural gene, A,, came under the
control of the two-element system of which Dotted (Dt) is the regulator
(Laughnan, 1955; Sarma, 1956, 1961).

This report will describe the mode of operation of the elements that com-
pose a single control system in maize. It is not the purpose of the paper to
present the evidence for the statements that will be made here but rather to
indicate some of the resemblances between the systems in bacteria and
those in maize. The Suppressor-mutator control system in maize has been
chosen because it illustrates these resemblances more directly than do
other examined systems in maize.

Five independent inceptions of control of gene action by the Suppressor-
mutator (Spm) control system have been recognized in the Cold Spring Har-
bor cultures. Three of them occurred when the ‘‘operator’’ element of this
system was inserted at the locus of A; in chromosome 3. These three cases
are designated a,"~1, a,™~2, and a,™-5. The symbols m-1, m-2, and m-5
refer to the order in time of inception of control of gene action at A, by this
system (a,"~3 and a,™~4 refer to inceptions of control of gene action at A;
by the Ds-Ac control system). A fourth case occurred at the A locus in
chromosome 5 (designated a,”~1), and a fifth at the Wx locus in chromo-
some 9 (designated wx”—8). (Both A, and A> are associated with anctho-
cyanin pigment formation in plant and kernel. Wx is associated with produc-
tion amylose in the pollen grain and in the endosperm of the kernel.) Two
independently located elements are primarily responsible for control of gene
action at am}, a,™-5, ay™-1, and wx™-8, One controlling element, com-
268 THE AMERICAN NATURALIST

Parable to the operator, resides at the locus of the gene and directly con-
trols its type of action. The other element, comparable to the regulator, is
Spm, to which the ‘‘operator’’ element at the locus of the gene responds in
accordance with the type of activity of Spm and the changes in this activity.
Preliminary evidence suggests that Spm resides close to the A, locus in the
case of a;"~2. It is possible that here the “operator”? and “‘regulator’’ ele-
ments of the system are located adjacent or close to each other. If this
proves to be true, a,”~? will resemble in its organization and its behavior
one of the cases examined in studies of the Ds-Ac system. (See discussion
of bz™~2 in McClintock, 1956.) ,

Basically, the mode of control of gene action by the Spm system is rela-
tively easy to comprehend. However, there are some conditions that compli-
cate its analysis and so obscure its basic Simplicity. They arise from
(1) alterations at the locus of the gene (termed ‘‘altered states’’ of the gene
locus), induced by the controlling element there residing, that modify sub-
sequent expression of the gene, both in the presence and in the absence of
Spm; (2) modifications of Spm itself, expressed by altered degrees in
strength of its action, or by cyclically occurring change in phase of its ac-
tivity—from active to inactive and back to active; and (3) the action of an
independently located, transposible Modifier element that alters the expres-
sion of some of the states of the gene locus in a predictable manner, but
only when Spm is also present and in its active phase. Each of these con-
ditions will be considered in turn. The discussion will apply to those cases
in which the two elements that are basically concerned in control of gene
action are independently located, one being at the locus of the gene, the
other, Spm, being located elsewhere.

THE CLASS I AND CLASS II STATES OF A GENE LOCUS UNDER
THE CONTROL OF THE Spm SYSTEM

There are two main categories of State, designated class I and class II.
Because the class II states behave in a simple manner, they will be con-
sidered first. In the Presence of a fully active Spm, no gene action is ex-
pressed. If Spm is removed by somatic transposition, or by meiotic segrega-
tion, or if it enters its inactive phase in a cell of the plant or the kernel,
gene action is expressed. The degree of expression serves to distinguish
between different members of the class II States. An apparently full, or near
full gene expression characterizes some class Il states whereas a much re-
duced expression characterizes others. With the class II states, Spm serves
as the ‘‘regulator’’ of action of the ‘‘structural”’ gene, causing it to be
“‘turned on’? and ‘‘turned off” through the direct mediation of the “‘operator”’
element residing at the locus of the “structural” gene. It should be empha-
sized that this turning on and turning off of gene action is not accompanied
by any modification that permanently alters the structure of the gene locus,
as may occur with the class I states, However, the class II states originate
from the class I states, which will now be described.
GENE CONTROL IN MAIZE AND BACTERIA 269

Because of the variety of expressions that may. be produced by class I
states, they appear to be far more complex than the class II states. Regard-
less of the degrees of difference in expression, all class I states exhibit
the same basic pattern of behavior. With respect to any one gene under the
control of the Spm system, all the many different class I states that have
been isolated trace their origin to the class I state that was produced, ini-
tially, when the ‘‘operator’’ element became associated with the structural
gene. Each class I state is distinguished not only by its behavior pattern
in. the presence of active Spm, but also by the type of gene action that it
gives rise to in the absence of Spm (or in the presence of Spm in an inactive
phase). When a fully active Spm is present, all gene action is suppressed
until, in a cell of the plant or of the kernel, a modification is instigated at
the locus of the gene by the ‘‘operator’’ element there residing. Each class
I state is distinguished by a particular type of consequence of such modifi-
cations. There are two main consequences: production of a mutant, which is
thereafter stable in the presence of active Spm; or production of a new state,
either class I or class II. The time during development of a plant or kernel
when these modifications occur, the number of cells in which they occur at.
any one stage in development, and the particular types of consequence,
serve to characterize a particular class I state. For example, with some
states, such modifications at the locus of the gene may occur early in de-
velopment, but each state is distinguished from others by the type of con-
sequence of these modifications. One such state may give rise to two main
types of stable mutants, those that express high levels of gene action and
those that give the null expression; and these two types of mutants are pro-
duced at constant relative rates. With another such state, the early-occurring
mutations result only in mutants that express low levels of gene action.
Some class I states give rise to many new states whereas others produce
few new states.

There is a group of class I states characterized by the fact that all
mutation-inducing events occur late in development of plant or kernel. The
states in this group may be distinguished from one another by differences in
the number of phenotypically distinguishable mutant areas that are produced,
and also by the levels of gene action these mutant areas exhibit. These
states are particularly useful for many studies. Mutation-inducing events
occur so late in development that germinal mutations may not be encountered
or ate encountered only rarely. Thus, such states are preserved, unaltered,
through generations of plants, even when Spm is present and fully active in
the plants.

In the absence of Spm all but one of the many class I states that have
been isolated express some degree of gene action. The level of action may
be low with some states, intermediate with others, and even quite high with
still others. As long as Spm is absent (or is present in its inactive phase),
the particular type of gene action expressed by any one state is constant
and may be maintained unaltered from generation to generation. Also, in the
270 THE AMERICAN NATURALIST

absence of an active Spm, any state behaves as a stable allele of any other
state. However, no relation has been observed between the type of gene ex-
pression that a class I state exhibits in the absence of Spm and the types of
mutation it produces in its presence.

If, in a plant having an active Spm, a different class I state is carried in
each homologue, each state reacts to Spm in its own individual manner, and
each may be recovered in the progeny, unaltered by its association in the
same nucleus with the other state. Again, if two different gene loci, each
under the control of the Spm system, are present in a plant or a kernel—for
example, a,”—1 and wx™~-8—each responds to Spm in its own characteristic
manner according to its state. It may be added, also, that in plants or ker-
nels having two such states each state responds directly to a somatically
occurring change in action of Spm. Each depicts this change in the expected
manner, in accordance with its state (see below).

TYPES OF CHANGE IN Spm

Spm itself undergoes modification. After a modification has occurred, the
Spm exhibiting it may be isolated and further examined. The modifications
may result in one of several different types of change in Spm action or be-
havior. One type affects the time during development of the plant when
transposition of Spm will occur. Some isolates of Spm undergo transposition
mainly early in development. Others undergo transposition mainly late in
development, with only an occasional occurrence in young tissues. Several
isolates have been obtained that rarely undergo transposition at times dur-
ing development that will resule in che appearance of gametes in which Spm
occupies a new location in the chromosome complement; and one isolate
has not yet given any evidence of transposition.

One conspicuous type of change undergone by Spm results in a weakening
of its capacity to effect mutation with the class I states of those gene loci
that have come under the control of the Spm system. The designation Spmv
is used to symbolize this type of alteration of Spm (McClintock, 1957). In
this section, the symbol SpmS will henceforth be used to designate an Spm
expressing full activity. A newly arising Spm may be recognized readily.
With the class I states, its presence is made evident by a pronounced delay
in time of occurrence of mutation at the locus of the gene, and also in a pro-
nounced reduction in frequency of occurrence of such mutation. If the class
I state is one that gives only late-occurring mutations with Spms, then with
Spm’ only a few very late-occurring mutations will be produced, and some-
times none at all. If the class I state is one that gives many early-occurring
mutations with Spm‘, then Spm’ will delay che cime of occurrence of mu-
tation until the late stages of development of a tissue. Only small areas
exhibiting the mutant phenotype will appear in plant and kernel. The re-
sponse of any one class I state to any one isolate of Spmw is quite
predictable. ,

Each Spm arises from an SpmS as the consequence of a single event oc-
curring within a cell. If it takes place early in plant development, all the
GENE CONTROL IN MAIZE AND BACTERIA 271

cells producing an ear of the plant may be descendants of the cell in which
it occurred, and thus all carry the newly produced Spm. Or, the descendent
cells may contribute only to a part of the ear, and the newly produced Spm”
be evident only in the kernels within a sector derived from these cells. In
either case, the Spm’ may be isolated from those kernels that carry it.
There are different types of Spm”, distinguishable from each other by sev-
eral criteria. They differ in degree of weakening of the capacity to induce
mutation at the loci of genes controlled by the Spm system, in frequency of
occurrence of transposition and the time of its occurrence during develop-
ment, and in stability of the Spm’ expression. In the last-named respect,
differences between Spm” isolates of independent origin are conspicuous.
Some isolates are highly stable whereas others undergo frequent return to
Spms. It is again evident that change in action of Spm, this time from Spm”
to Spms, is effected by a single event occurring in an individual cell.

If both an Spm’ and an Spm are present in the same plant, Spm’ is
dominant. However, the Spm’ in such plants may be recovered in their
progeny, with its type of action unaltered by previous association in the
same nucleus with Spms.

It should be emphasized that no modifications resulting in mutation or
change in state will occur at the locus of a gene that is under the control of
the Spm system (that is, with the class I states of the gene locus), unless
all gene action has first been suppressed by Spm in the ancestor cells,
whether by Spm’ or by Spm5. In other words, suppression of gene action by
Spm must precede the mutation-inducing event. This fact is especially well
illustrated in plants that have an Spm element that is undergoing change in

phase of its activity during the development of plant or kernel, as described
in the next section.

CYCLICALLY OCCURRING REVERSALS OF PHASE OF ACTIVITY OF Spm

One of the most interesting and theoretically important types of expres-
sion of Spm consists in the sequentially occurring reversals in phase of its
activity—from active to inactive and back to active (McClintock, 1958,
1959). Each such change in phase results from an event occurring in an in-
dividual cell of the plant or kernel. The effect produced by the change is
then exhibited in the descendants of this cell, if either a class I or a class
II state of a gene locus under the control of this Spm system is also present
to register it. Following such a reversal of phase, the duration of the par-
ticular phase may be long, continuing unaltered through many cell or even
plant generations, or it may be short, reversal occurring again in a number
of cells only a relatively few cell generations removed from that which ini-
tiated the preceding phase. Control of duration of a particular phase appears
to be associated with the event that produces the particular reversal of
phase. By selective methods it has been possible to isolate Spm displaying

either a long duration of an active phase or a long duration of an inactive
phase.
272 THE AMERICAN NATURALIST

The phenotypes appearing in mature plants and kernels as the conse-
quence of phase reversal of Spm may be very complex. The degree of com-
plexity depends on the state of a gene locus that is present in the plant or
kernel, on the number of reversals of phase of Spm that occur, and also on
the times of their occurrence during development of a tissue. If an inactive
Spm is present, initially, in a plant or kernel, along with either a class I or
a class II state of a gene locus, no evidence of the presence of this Spm
will appear in either the plant or the kernel unless reversal of phase occurs
in one or more cells during development. After Spm is reactivated in an in-
dividual cell, its presence is revealed in the descendants of that cell. When
a Class JI state is used as the indicator, reactivation of Spm is made evi-
dent by suppression of gene action in these cells. Should subsequent re-
versal of phase occur in some of the descendent cells, then gene action will
be evidenced in their descendent cells. With the class II states, then, re-
versals of phase of activity of Spm merely effect a ‘turning on’’ and ‘‘turn-
ing off’’ of gene action; and with the class II states of those gene loci that
are associated with the production of anthocyanin pigment, both in plant and
in kernel, the alternating cycles of phase of activity of Spm are registered
with great clarity. No pigment appears in cells in which Spm is in its active
phase, and pigment appears in those cells in which it is in its inactive
phase. It may be pointed out here that the type of control of gene action,
just described, resembles that associated with phase variation in Salmonella
in which the system of chromosomal elements responsible for control of gene
action likewise has been identified (Lederberg and Iino, 1956; Iino and
Lederberg, 1957, 1958; Iino, 1959, 1960).

The response of the class I states to reversal of phase of activity of Spm
is basically the same as that of the class II states, but the types of pheno-
typic expression of the gene and the various different patterns of expression
that may appear in an individual plant or kernel can be very complex. This
is because mutation-inducing events may occur at the gene locus in some
cells when their Spm is in its active phase. The pattern of mutant areas
that may appear in a sector of the plant or kernel, after a change in phase of
Spm from inactive to active, will depend upon the developmental stage of the
tissue when the reversal occurs. An example will illustrate this. If a plant
or kernel starts development with an active Spm having a long duration of
the active phase, and also a class I state of the gene locus that responds
to it by producing a number of early-occurring mutations, then large areas,
each exhibiting a mutant. phenotype, will be present in the mature plant or
kernel. Each such area reflects an early-occurring ‘“‘operator’’-induced modi-
fication at the locus of the gene. If, however, development commences with
Spm in its inactive phase, no mutations may occur at the gene locus having
this class I state unless and until a reversal of phase of activity of Spm oc-
curs, If reversal takes place in a cell rather late in the development of a
tissue, suppression of gene action will be effected in the descendants of
that cell. However, mutation-inducing events may occur in some of these
descendent cells and, by necessity, all these mutations will arise in cells
GENE CONTROL IN MAIZE AND BACTERIA 273

of a tissue that is approaching maturity. Consequently, the areas that can
exhibit a mutant phenotype must be small. In other words, the size of the
mutant areas will depend upon the stage of development of a tissue when
Spm reverts to its active phase. It is evident, then, that complex and often
irregular patterns of gene expression may be exhibiced by a plant or by a
kernel carrying a particular class I state when its Spm is undergoing fre-
quent reversal of phase of activity. Different patterns of gene expression
may be observed in different areas of the same plant or kernel. These pat-
terns reflect the time of occurrence of reversal of phase of activity of Spm
and also the number of such reversals.

Evidence has been obtained to indicate that inactivation of Spm, as de-
scribed above, is not associated with a complete blocking of its functional
capacity but rather with some change affecting its mode of functioning, such
as an altered form of its product. This was made evident, initially, in plants
and kernels having an inactive Spm characterized by a long duration of inac-
tivity, and also an active Spm undergoing frequent reversal of phase during
development. If only the latter Spm were present, a class II state would
register each reversal of phase by showing no evidence of gene action in
those cells in which it was active and by exhibiting gene action in those
cells in which it was inactive. When two active Spm elements are present,
initially, a class II state registers reversal of phase only when it occurs to
both Spm elements, either simultaneously in an individual cell, or succes-
sively (that is, affecting one Spm in one cell and the other Spm in a de-
scendant of that cell). Thus, in either plant or kernel, both the number and
the size of areas exhibiting gene action will differ according to the number
of active Spm elements that were present initially. From these patterns, it
is often possible to deduce the number of Spm elements that are present ina
plant or kernel.

It was anticipated that combination of an inactive Spm, having a long du-
ration of the inactive phase, with an initially active Spm in a plant or kernel
carrying a class II state would give rise to a phenotype resembling that pro-
duced when only one active Spm is present initially. This assumption proved
to be incorrect. Instead, it was found that this combination produced a
phenotype resembling the one that appears when two active Spm elements
are initially present in a plant or when three active Spm elements are ini-
tially present in a kernel. However, the pattern produced by the areas that
exhibit gene action (no active Spm in them) is much more uniform, and this
is particularly well illustrated in the aleurone layer of kernels whose endo-
sperms receive two inactive Spm elements from the female parent and one
initially active Spm from the male parent. All areas exhibiting gene action
are small, and they are evenly distributed over the aleurone layer. That this
pattern is not produced by reversal of phase of the inactive Spm, brought
about by association in the same nucleus with an active Spm, is made evi-
dent when progeny of plants having an initially active Spm and the described
inactive Spm are examined. The inactive Spm is recovered with its phase
quite unaltered. Also, it appears in the expected proportions of the progeny
274 THE AMERICAN NATURALIST

in accordance with the type of testcross that has been made to determine
this. In order to be certain that the inactive Spm appearing in the progeny
was the same as'that which had been combined with the active Spm in the
zygote produced from the initial cross, the relative locations of the two Spm
elements in the chromosome complement had to be known in advance of the
initial cross. Also, the location of each had to be determined in the indi-
vidual progeny.

A number of tests had been made to observe the effects produced on either
the class I or the class II states by bringing together in a zygote nucleus,
or in a primary endosperm nucleus, an inactive and an active Spm. The ef-
fects produced in all such tests conformed with that described above. The
inactive Spm proved not to be totally inactive, although it was quite ineffec-
tive by itself. This evidence does not preclude the possibility or the proba-
bility that some modifications of Spm may result in its total inactivation.

THE MODIFIER ELEMENT IN THE Spm SYSTEM

A transposible element that serves to increase the frequency of occur-
rence of mutation-inducing events with some of the class I states of a,™—1
first appeared in only one of many a,™~1, Spm-catrying kernels on an ear
and on only one of several ears produced by an a,"~1/a,, Spm-carrying
plant. This kernel exhibited a marked increase in mutation frequency in
comparison with that exhibited by the other a,;”~1, Spm-carrying kernels on
the ear. The class I state of a,”~1 that was Present in the ear-bearing plant
was one that undergoes only late-occurring mutations in the presence of an
active Spm. No change in this state had been observed to occur in many
tests conducted with it over a number of plant generations. The presence of
a Modifier element, which was responsible for the marked increase in muta-
tion frequency in the exceptional kernel, was revealed in tests conducted
with the plant derived from this kernel. Subsequently, the effects produced
by this Modifier on the expression of other class I states of ay
vestigated. Study of its effects was confined to ay
low the following conclusions to be drawn:

(1) The presence of the Modifier can be detected only when Spm also is
Present in the chromosome complement and only when it is in its active
phase. Under these circumstances, the Modifier effects a marked increase
in frequency of mutation to stable alleles with some of the class I states,
but does not modify the time of occurrence of such mutation. Also, the rate
of increase in frequency-of mutation is proportional to that produced by the
state in the absence of the Modifier. However, if the state is one that pro-
duces very many mutations with Spm alone, the Modifier does not effect a
measurable increase in mutation rate (McClintock, 1958),

(2) When the Modifier is present, the same phenotype is produced with
Spm” as with Spms. Thus, plants and kernels that have only Spm’ and the
Modifier are not distinguishable in phenotype from those that have Spms and
the Modifier. However, the presence of either one or the other type of Spm
may be determined by means of progeny tests. Individuals carrying Spm but

m-1 were in-
m-l, but the results al-
GENE CONTROL IN MAIZE AND BACTERIA 275

no Modifier appear in the progeny, and the type of Spm ia them is made
evident.

(3) The Modifier element is transposible. A number of early-occurring
transpositions of it were detected. Its transposition to and away from loca-
tions in the chromosome complement close to marked gene loci were exam-
ined (McClintock, 1958).

The Modifier element acts as if it could complement both the ‘‘regulator”’
element, Spm, and the element of this system that is at the locus of the
gene. It complements Spm in that in its presence a weakly acting Spm
(Spm) is as effective as a fully active Spm (Spms). It complements the
element at the A; locus in that in its presence a class I state that gives
relatively few mutations with Spm* alone can mimic another state that gives
many more mutations with Spm alone.

DISCUSSION AND SUMMARY

Although the mode of operation of the Spm system of control of gene ac-
tion, as outlined above, may appear to be complex, it is evident, neverthe-
less, that the diverse gene expressions that it may produce stem from one
basic mechanism of action and response of the component elements of the
system. The action of the Spm element resembles that of the regulator
‘tgene’’ in bacteria. It may well be that Spm produces a specific repressor
substance to which the element of the system at the locus of the gene, the
‘toperator’’ element, responds by ‘‘turning off’’ gene action. Suppression of
gene action requires the presence of this operator element at the locus of
the gene. No suppression occurs when an operator element belonging to an-
other system is present at the gene locus, or when the specific operator ele-
ment of this system is transposed away from the gene locus. In other words,
if Spm produces a specific repressor substance, then the operator element at
the locus of the gene responds only to this specific repressor and to no
other. The same principle would apply to all the two-element control sys-
tems investigated so far in maize; and in this respect they resemble the two-
element control systems in bacteria.

In bacteria, both the operator and the regulator element undergo mutation.
The mutations arise from single events, and some of them are reversible.
The same applies to the controlling elements in maize. Each can undergo
mutation, and each such mutation is produced by a single event. Also, some
of them are reversible.

In bacteria, most of the control systems that are subject to analysis ef-
fect control of production of specific enzymes in response to certain
changes in the cellular environment. A “turning on’ and ‘‘turning off”’ of
gene action constitute an efficient means of control of production of en-
zymes in response to changes in intracellular environment. Mutations, such
as those that occur with the class I states described above, would effect a
differentiation. Some of them could be lethal or could result in competitive
disadvantage for unicellular organisms. In higher organisms, such muta-
tions, occurring in somatic tissues, need not be lethal or disadvantageous
276 THE AMERICAN NATURALIST

and, indeed, may be required. Specific types of mutation occurring at given
times during development and produced by a control system, such as the Spm
system, may effect tissue differentiation along certain paths. However, as
emphasized above, the Spm system can operate in either way-—in a manner
similar to that exhibited in bacteria, or in a manner that accomplishes a per-
manent and specific type of change in gene action.

The class II states of gene loci under the control of the Spm system best
illustrate the similarities in mode of operation of the bacterial and the maize
systems. The ‘‘operator’’ element is fixed in location. No transpositions
of it away from the gene locus occur, nor does it effect mutation at the gene
locus. Its behavior is much the same as that of the operator ‘‘gene’’ in bac-
teria. If Spm produces a specific repressor substance, then the operator
element responds to this by ‘‘turning off’ gene action. If the repressor sub-
stance is not produced, or if its structure is modified by mutations that oc-
cur to Spm, then gene action is ‘‘turned on.’”’ A class II state, with its
operator element fixed in position and an Spm element that also is fixed in
position, gives rise to a system of ccntrol of gene action in maize that simu-
lates in its mode of operation some of the described systems in bacteria.
As stated earlier, cases of effective fixation of Spm ata specific locus have
been found. .

Study of the Spm system has shown that a relatively simple system of con-
trol of gene action may be derived from one that otiginally expressed a
seemingly complex pattern of such control and it may be no coincidence that
this simple system resembles those recently discovered in bacteria and in
phage. It is expected that such a basic mechanism of control of gene action
will be operative in all organisms. In higher organisms, lack of means of
identifying the components of a control system of this type may be respon-
sible for delay in recognition of their general prevalence, even though there
is much genetic and cytological evidence to indicate that control systems
do exist. It is anticipated, however, that control systems exhibiting more
complex levels of integration will be found in the higher organisms.

LITERATURE CITED

Brink, R. A., 1958, Mutable loci and development of the organism. J. Cellu-
lar Comp. Physiol. (Suppl. 1): 169-196,

1960, Paramutation and chromosome organization. Quart. Rev. Biol. 35:
120~137.

Buttin, G., F, Jacob and J. Monod, 1960, Synthése constitutive de galacto-
kinase consécutive au développment des bactériophage A chez Es-
cherichia coli K 12. Compt. rend. 250: 2471-2473,

Jacob, F., 1960, Genetic control of viral function. The Harvey Lectures,
Ser. 54: 1-39.

Jacob, F., and J. Monod, 1959, Génes de structure et génes de régulation
dans la biosynthése des protéins. Compt. rend. 249: 1282-1284.

1961, Genetic regulatory mechanisms in the synthesis of proteins. J.

Mol. Biol. 2: 318-356.
GENE CONTROL IN MAIZE AND BACTERIA 277

Jacob, F., D. Perrin, C. Sanchez and J. Monod, 1960, L’opéron: groupe de
génes 4 expression coordonnée par un opérateur. Compt. rend. 250:
1727-1729.

Jacob, F., P. Schaeffer and E. L. Wollman, 1960, Episomic elements in bac-
teria. Tenth Symp. Soc. Gen. Microbiol.: 67-91.

lino, T., 1959, Curly flagellar mutants in Salmonella. Natl. Inst. Genet.
(Japan) Ann. Rept. 9: 95. .

1960, Unequal recombination in Salmonella. Natl. Inst. Genet. (Japan)
Anon. Rept. 10: 112-113.
lino, T., and J. Lederberg, 1957a, Transductional analysis of phase varia-
tion in Salmonella. Nati. Inst. Genet. (Japan) Ann. Rept. 7: 89-91.
1957b, An analysis of monophase variation in Salmonella. Genetics 42:
378.
1958, Transductional analysis of monophasic types of Salmonella. Natl.
Inst. Genet. (Japan) Ann. Rept. 8: 101-102.

Laughnan, J. R., 1955, Structural and functional bases for the action of the
A alleles in maize. Am. Naturalist 89: 91-103.

Lederberg, J., and T. lino, 1956, Phase variation in Salmonella. Genetics
41; 743-757.

McClintock, B., 1953, Induction of instability at selected loci in maize. Ge-
netics 38: 579-599.

1956a, Intranuclear systems controlling gene action and mutation.
Brookhaven Symp. Biol. 8: 58-74.

1956b, Controlling elements and the gene. Cold Spring Harbor Symp.
Quant. Biol. 21: 197-216.

1956c, Mutation in maize. Carnegie Inst. Wash. Yrb. No. 55: 323-332.

1957, Genetic and cytological studies of maize. Carnegie Inst. Wash.
Yrb. No. 56: 393-401.

1958, The Suppressor-mutator system of control of gene action in maize.
Carnegie Inst. Wash. Yrb. No. 57: 415-429.

1959, Genetic and cytological studies of maize. Carnegie Inst. Wash.
Yrb. No. 58: 452-456.

Peterson, P. A., 1960, The pale green mutable system in maize. Genetics
45: 115-133.

Richter, A., 1961, Attachment of wild type F factor to a specific chromo-
somal region in a variant strain of Escherichia coli K 12: The phe-
nomenon of episomic alternation. Genet. Res. 2 (in press).

Sarma, M. S., 1959, A test of the mutational hypothesis for the origin of the
noncrossover alpha derivatives from Ab:P and A>: Ec. Maize Ge-
net. Cooperation News Letter 33: 44-48.

1961, Colorless components of the Ab:Ec and A’:P complexes in
maize. Maize Genet. Cooperation News Letter 35: 47-50.
Yarmolinsky, M. B., and H. Wiesmeyer, 1960, Regulation of coliphage lambda
of the expression of the capacity to synthesize a sequence of host

enzymes. Proc. Natl. Acad. Sci. U.S. 46: 1626-1645.