ASPECTS OF GENE REGULATION IN MAIZE
Barbara McClintock

Investigations with maize during the
past year were designed with several
objectives in mind. They ranged from the
routine but necessary determination of
whether or not a particular gene had come
under the control of one of the known
gene-control systems, or some yet unde-
fined system, to a more basic examination
of the extent and diversity of gene control
during development that may be ex-
pressed by a single system of controlling
elements.

It has long been apparent that a single
gene-control system in maize can effect
a wide range of expression with respect to
time and type of gene action during
development. It has been possible to
relate some aspects of this diversity to
modifications arising solely within one or
the other element of a control system.
There is now a considerable body of
evidence to indicate the presence in higher
organisms of different classes of regulatory
mechanisms that must integrate with one
another in the control of gene expression.

Some of these mechanisms effect control
of the appearance of the end product of
a particular biosynthetic pathway in only
some cells of a tissue, even though all cells

_ are known to be competent to express the

end product. Their action is reflected in
the appearance of distinct patterns of
distribution of the end product within
different tissues or different parts of an
organism.

It is known from early genetic studies
that regulation of such patterns may
sometimes reside in a single chromosomal
locus. The distinctly different distribu-
tions of pigment in the elytra of the lady
beetle, investigated by C. C. Tan
(Genetics, 1946), and the distribution of
bristles in Drosophila, referred to as ‘‘step
allelomorphism” and discovered many
years ago, are evidences of genetic ele-
ments that must incorporate a mechanism
able to control’ the activity of particular
genes within certain cells of a developing
tissue, each allele of the locus being
responsible for one pattern. In these
GENETICS RESEARCH UNIT

examples, the mechanism comprised
within one allele operates independently;
whenever two different alleles are present
in the nuclei of a developing tissue, each
functions to produce its own predeter-
mined distribution of gene product, and
the resulting patterns overlap. The
isolation of such alleles demonstrates that
mutations affecting only the control
mechanisms must occur.

Studies with maize have indicated that
controlling elements may be solely respon-
sible for this kind of regulation of gene
action during development, and that the
alleles represent mutants of these ele-
ments. The controlling elements behave
as independent genetic agents, controlling
not only the distribution of a gene product
in a developing tissue but also its level or
type. Because they regulate the time of
expression of particular sets of genes, they
serve as “genetic clocks.” The different
alleles of an element may be said to
represent different “‘settings’ of the
clock. They arise by mutation, and each
mutant is recognized by the marked
change it effects in the time or type of
gene action during development of a
tissue. Plate 1A illustrates the distinctive-
ness of such mutants. In the past these
mutations have been referred to as

593

“changes in state.” Each state retains its
characteristic mode of control of gene
action through plant generations unless
a subsequent mutation again alters the
control mechanism.

The “operator” and “regulator” ele-
ments of a control system may function
together as component parts of the
clocklike mechanism. Combinations, in a
zygote or primary endosperm nucleus, of
different states of the operator of a system
with different states of the regulator can
produce a series of widely divergent
patterns and types of gene expression in
the mature tissue. The controlling ele-
ments of a single system, therefore, are
exceptionally versatile genetic agents in
that they can give rise to a variety of
different expressions of a gene during
development.

That the controlling elements in maize
are even more versatile than was previ-
ously suspected will be illustrated in the
following sections. Opportunity will be
taken first, however, to indicate (table 1)
the different instances, examined in the
Cold Spring Harbor cultures, in which 7
genes, distributed among 4 different
chromosomes of the maize complement,
have come under the control of either the
Ac (Activator) or the Spm (Suppressor-

TABLE 1. Examples Studied in the Cold Spring Harbor Cultures of Initiation of
Gene Control by the Ac and Spm Systems

Symbol for Locus after Inception of Control by the Ac or Spm System

Gene Locus

 

Controlled by the Ac System

Chromosome 3

Controlled by the Spm System

Ai a3, ay™-4 aim), a,m2,* ay™-5*
Chromosome 4

Cy cm1* com-2
Chromosome 5

Ao aym-4 amt, amb
Chromosome 9

C1 c™-1, cam?, cam-3, eym4 cyn-5*

Sh shi™}, shym-2

Ba bai}, bey™-2,* bzy-4

We war), wan, wars, wrrs

was, wart* wom

 

* Both the operator element and the regulator element were initially at the locus of the gene.
594

mutator) system. Some of the same genes,
as well as others not listed here, have
come under the control of other systems,
whose modes of operation have not yet
been adequately explored.

Parameters of Regulation of Gene
Action by the Spm System

A type of Spm control examined during
the past year illustrates one way in which
a “genetic clock’”’ may be set recurrently.
The effects of the settings are expressed
in the cells of the aleurone layer of the
kernels through the control each setting
exerts over intensity levels and patterns
of distribution of anthocyanin pigment.
They are observed in one genetic class of
kernels on ears produced by crosses of
plants carrying Spm and the second of the
two modified states of ai"? discussed in
Year Book 62, where the origin of the
state and the initial tests made with it
were outlined.

Early studies of this state (state 7977B)
indicated the absence of an active Spm
element at the locus of Ai although one
was present at that locus in the original
isolate of ai". (The A, gene in chromo-
some 3 is involved in the production of
anthocyanin pigment in plant and ker-
nel.) The initial plant carrying the
modified state had no active Spm at any
other location in the chromosome comple-
ment, but an active Spm was introduced
by an appropriate cross conducted with
one of the two ears of the plant. In
kernels of this ear that received from the
female parent the modified state of ai"?
and from the male parent the standard
recessive allele (a1) and also Spm, the
expression of gene action resembled that
engendered by the original state of a)".
(See legends, pls. 1B, 2A and B.) Year
Book 62 described the results of test-
crosses made with plants derived from the
a,™-*—~carrying kernels on this ear that had
received Spm and those that had not
received Spm. The description placed
emphasis on the appearance of a previ-
ously unobserved synergistic mode of

CARNEGIE INSTITUTION

formation of anthocyanin pigment in the
cells of the aleurone layer of some kernels
on ears produced by crosses conducted
with the plants that had Spm in their
nuclei. The phenotypes of kernels in this
class did not differ on ears produced by
reciprocal crosses conducted with the
plants. In other words, there was no effect
of dose of ai"? on intensity levels or
patterns of distribution of anthocyanin in
the aleurone layer. Although the early
evidence indicated that the kernels with
pronounced expressions of this unusual
phenotype did not contain an active Spm,
no such kernels appeared in similar
crosses made with sister plants that had
no active Spm in their nuclei. Neither did
they appear on ears produced by recip-
rocal crosses of these sister plants with
plants that were homozygous for standard
a; and carried one or more active Spm
elements. These initial observations indi-
cated that Spm, present in a plant, is
involved in controlling phenotypic expres-
sion in progeny kernels that do not receive
Spm. Experiments conducted this year
not only have confirmed that effect but
also have revealed still another aspect of
regulation of gene action during develop-
ment.

Among kernels exhibiting the unusual
phenotypes, both the patterns of antho-
cyanin distribution and the range of
intensities of pigment among cells of the
aleurone layer vary, and are distinctive in
each kernel. The maximum pigment
intensities range from very high in some
kernels to faint in others. The variations
are illustrated by several of the kernels
on the ear shown in plate 2A, and by 5
of the 6 kernels in plate 2B. It is evident
that the phenotype of each such kernel
must arise through some form of control
of anthocyanin production and distribu-
tion in individual cells during endosperm
development, and that the pattern ex-
pressed in a mature kernel reflects a
“presetting’”’ of the responsible compo-
nent. It is also evident that this setting
must occur before the three haploid nuclei
(two from the female and one from the
GENETICS RESEARCH UNIT

male) fuse to form the primary endo-
sperm nucleus. Since Spm is not present
in these nuclei, but must be present in the
male or female parent plant carrying state
7977B of ai”? in order that pronounced
expressions of the unique phenotype may
appear, the mechanism responsible for the
setting must be activated before Spm is
removed from spore nuclei as a conse-
quence of the meiotic mitoses. Evidence
will be presented below that such
activation is independent of the develop-
mental stage and requires only the initial
presence of an active Spm in the plant.

This year, the types of experiments
outlined in Year Book 62 were repeated
on a much larger scale, not only with
state 7977B but also with another state
of a"? (7995) that resembles it in some
aspects of expression. In addition, studies
were made to determine the effects of
inactive Spm and of Spm” as well as Spm'
in the progeny kernels of plants carrying
state 7977B. Plants were grown from a
number of kernels that exhibited the
unusual phenotypes, and they in turn
were subjected to several kinds of test.
The results of these studies are summa-
rized below.

Kernels exhibiting pronounced ‘pre-
set”? patterns, similar to those illustrated
in plate 2A and B, appeared on testcross
ears of plants carrying an active Spm,
which might be either Spm* or Spm”.
(The distinction between the effects pro-
duced by Spm* and Spm” on the original
state of a,:”-? is shown in pl. 1B.) All
together, 84 plants carrying state 7977B
and 28 plants carrying state 7995 were
examined. When these plants were
crossed with plants that were homozygous
for the standard a, allele and had no Spm,
several of them produced ears containing
a sector that had developed from a cell
in which Spm had been removed or
inactivated. Within these sectors, never-
theless, some kernels exhibited the pro-
nounced preset patterns of anthocyanin
distribution. In one plant, the ear
produced by one tiller did not carry active
Spm, although the ears on the main stalk

595

and on a second tiller did. On the ear in
which Spm was absent, kernels showing
pronounced preset patterns appeared.
The range in types of pattern and in
maximum pigment intensities among
these kernels, and among those within the
no-Spm sectors of the sectored ears, was
as wide as that seen among kernels on
the ears of plants that showed no evidence
of somatic loss or inactivation of Spm.
No kernels of the exceptional type
appeared, however, on testcross ears
produced by 9 plants carrying state
7977B that commenced development
with an inactive Spm, except within
sectors in which Spm had changed to its
active phase. These tests made it evident
that an active Spm must be present in a
plant, at least initially, to promote the
mechanism that is responsible for pro-
nounced preset, synergistically produced
anthocyanin patterns in progeny kernels
not receiving Spm.

Plants were grown from 51 kernels
exhibiting such patterns, selected from
six ears produced by testcrosses conducted
with 3 plants having state 7977B of ai"?
in chromosome 3 and an active Spm in
one chromosome 5. Testcrosses designed
to detect presence or absence of Spm were
made with 43 of these 51 plants. (This
type of test has been outlined in previous
Year Books.) None of the 43 had an
active Spm element. Plants were also
grown from 7 kernels of similar type from
an ear of a plant carrying state 7995 of
a,""2 and two active Spm elements,
neither of which was linked with the a:"-?
locus. No evidence of active Spm was
found in tests conducted with 5 of the 7
plants.

In order to determine the kinds of
pigment patterns that might appear
among the kernel progeny of these 58
plants, each was subjected to additional
tests. In one test, at least one ear of each
plant was utilized as female parent in a
cross with a plant that was homozygous
for standard a; and had no active Spm.
(Reciprocal crosses were made with only
4 of the plants, but the results were quite
596

the same.) Ninety-two ears were pro-
duced by crosses of this type. With the
exception of a very few kernels on ears of
plants carrying state 7995, no phenotype
resembling that of the kernel from which
each plant arose reappeared among the
kernels on these ears. Instead, the a1-?-
carrying kernels were similar to those on
one ear, produced by the same type of
cross, on the original plant carrying state
7977B. It will be recalled that this plant
had no active Spm. Such kernels may be
nearly colorless; or they may have a few
darkly pigmented cells, or areas of faint
pigmentation, or both. This expression of
state 7977B has continued to reappear in
testcrosses conducted with three suc-
cessive generations of plants (99 plants
in all) into which no active Spm was
introduced. Pigment patterns such as
those shown in plate 2B have not ap-
peared among the progeny kernels.

A second kind of testcross was con-
ducted with some of the 58 plants derived
from the kernels with pronounced preset
patterns. One ear each on 24 of the 51
plants carrying state 7977B and on 3 of
the 7 plants carrying state 7995 was
utilized in a cross with a plant that was
homozygous for standard a; but had one
or more active Spm elements. Two classes
of a,"-2-carrying kernels appeared. Ker-
nels that did not receive Spm were nearly
colorless or had areas of faint pigmenta-
tion; their phenotypes were similar to
those just described. Kernels that re-
ceived an active Spm had phenotypes
resembling that evoked by the original
state of ai%%. (See, in pl. 2A, the Shz
kernels with many dots of deep pigment
in a lighter background, and in pl. 2B,
the leftmost kernel, upper row.) No
kernels like the other five shown in plate
2B appeared on these ears.

The 58 plants just described were
derived from kernels with pronounced
preset patterns, from ears whose kernel
types resembled those seen in plate 2A.
The photograph shows, in addition to the
kernels with preset patterns, two other
classes of Sh2 kernels: those having

CARNEGIE INSTITUTION

deeply pigmented spots on a lighter
pigmented background (Spm present),
and those that are colorless or nearly so
(no Spm). Plants were grown from both
types of kernels, selected from the same
ears as the kernels from which the 58
plants were derived. From the near-
colorless kernels, 28 plants carrying state
7977B and 10 plants carrying state 7995
were grown and tested. From the kernels
that had received Spm, 32 plants with
state 7977B and 19 with state 7995 were
grown and examined. The testcrosses
made with the 38 plants derived from the
near-colorless kernels were like those
conducted with the 58 plants grown from
kernels having pronounced preset pat-
terns. None of these plants carried an
active Spm. In crosses with plants
homozygous for a; and having no Spm,
none of the progeny kernels exhibited
patterns like those shown in plate 2B.

The a:”-2-carrying kernels were colorless

or nearly colorless. When the testcross
introduced an active Spm, the a,”"? locus
responded in the expected manner, giving
rise to deep spots in a lighter background.
In contrast, all 51 of the plants derived
from kernels containing Spm produced
some progeny kernels whose patterns of
anthocyanin distribution and intensity
levels resembled those shown in plate 2A
and B,

Another set of tests was carried out to
reconfirm the conclusion that Spm must
be present in a plant to initiate the
modifications responsible for pronounced
preset patterns in some of its no-Spm
kernel progeny. These tests were made
with second-generation progeny of the
original plant carrying state 7977B of
a,"-?, One ear of that plant had been
produced by a cross with a plant homozy-
gous for standard a, and having no active
Spm; and plants had been grown from
the a,"*-carrying kernels on that ear.
Some of these plants, in turn, were
crossed with plants that were homozygous
for a, but carried one or more active Spm
elements. Plants were then grown from
both Spm-carrying and Spm-less kernels
GENETICS RESEARCH UNIT

produced by three such crosses. Forty-
seven plants derived from Spm kernels
and 20 derived from no-Spm kernels were
crossed with plants that were homozygous
for a; and had either no active Spm or
one or more active Spm elements. Once
again, kernels -exhibiting the unique
phenotype appeared on ears of the Spm-
carrying plants but not on the ears of
plants lacking Spm.

Thus the results of experiments con-
ducted with three generations of plants
carrying state 7977B, and of the more
limited but still instructive experiments
with plants having state 7995, leave no
doubt about the role played by active
Spm in a plant in controlling the pheno-
types of its kernel progeny, both those
that receive Spm and those that do not.
It is also clear that Spm does not directly
control the presetting mechanism itself,
as was shown by the presence, in no-Spm
sectors on ears of several Spm-carrying
plants, of kernels with preset patterns.
The presetting event, unlike the events
that give rise to altered states, does not
modify the a," locus so as to alter the
potential for gene expression in a fixed
manner. Rather, it appears to involve a
release from restrictions, which permits
the gene-control mechanism to follow
paths that would otherwise be closed.
The meaning of “path” in this reference
may be clarified by consideration of the
distinctive phenotypes of the variegated
kernels on the two ears shown in plate 1A.

The cross that produced each of these
ears was made to test for the presence and
number of Spm elements in one of the
parent plants, which was homozygous for
the standard recessive of the gene whose
action is exhibited in the kernels on the
ear. The other parent had no active Spm
and was homozygous, in the upper ear,
for one particular state of a,"', in the
lower ear for a.”-!. It is known, as a result
of many years of testing, that each of
these states will respond to a fully active
Spm by producing a distinctive pattern
of anthocyanin distribution in kernels (as
shown in the photograph) as well as in

597

the plant. Each of many other isolated
states of a1”! and a.”"! also engenders its
own precise pattern of anthocyanin
distribution in the presence of a fully
active Spm. States of a1™! having expres-
sions similar to that of ao”! illustrated
here have been isolated, and also states
of a,” similar in expression to this state
of ay!,

These illustrations are ineluded to
emphasize a mode of differential regula-
tion of gene expression among the cells of
a tissue. Its control resides in a clocklike
mechanism built into the gene-associated
element of the control system. In the
illustrations given, the clock mechanism
was set in motion by the introduction of
an activating element, Spm‘, into the
primary endosperm nucleus. As long as
Spm? was present in the cells of the
developing endosperm, the clock con-
tinued to tick off precise events in certain
cells at particular stages of development.
The end result was a distinctive pattern
of distribution of the gene product in cells
of the mature tissue.

That cellular environment is not direct-
ly involved in control of these regulatory
processes was demonstrated, early in the
study of the Spm system, by crosses
between plants that carried different
states of ai}. The states were readily
distinguished from one another by their
markedly different responses to Spm.
Such crosses produced kernels carrying
two different states of a.*! as alleles of
the A, locus. For our purposes, the most
illustrative of these crosses combined one
state that responds to a fully active Spm
by producing both large and small areas
of light pigmentation with another state
that produces only dots of deep pigment,
in patterns resembling those seen in the
variegated kernels on the upper ear in
plate 1A. During the development of
kernels having both these states and a
single Spm*, each responded to the Spm
according to its own predetermined
ordering of the sequence of events. The
patterns and intensities resulting from the
individual responses were readily dis-
598

tinguishable in the aleurone layer, the
patterns overlapping each other. In other
words, each allele followed its own
directed path in the control of A: gene
action according to a built-in clock
mechanism, unaffected by that of the
other allele. The resemblance of this
phenomenon to the control exerted by
different alleles on the pattern of distri-
bution of pigment in the elytra of the lady
beetle, mentioned early in this report, is
striking.

In the examples just discussed, the
clock mechanism is permanently set to
follow a definite path during the develop-
ment of plant or kernel in successive
generations, unless a mutation occurs to
establish another clock setting. The
mechanism is contained in the gene-
associated operator element of a two-
element system. It may be slowed down,
however, or advanced to a maximum rate,
according to the activity of the regulator
element of the system. Mutations of the
regulator can modulate the timing of
responses but not their types, which are
controlled solely by the operator element.
Nevertheless, each allele of an operator
responds in the same manner to the
timing modulation induced by any one
mutant of the regulator. Thus, as was
stated earlier, both components of a
control system may be involved in the
performance of the clock mechanism.

In contrast to permanent settings of
the clock, which are responsible for the
origin and expression of the states dis-
cussed above, settings that are not
permanent may occur, as evidenced in
part by the observations regarding preset
patterns of anthocyanin pigmentation in
kernels, described in this report. It has
been shown that, although Spm is
required to initiate the mechanism re-
sponsible for these newly detected types
of setting, it need not be present in the
tissue cells that express the altered
settings. Nor do such settings perma-
nently alter the state of the operator,
which reverts in the subsequent plant
generation to that expressed when it was

CARNEGIE INSTITUTION

initially isolated. The impermanent set-
tings may result in a wide range of
distinctly different phenotypes, as illus-
trated by the no-Spm kernels shown in
plate 2A and B. The appearance of such
a range among the kernel progeny of a
single plant indicates that the setting
process must occur in a number of cells
and that its consequences are not the
same in all of them. Once a setting has
oceurred, however, the phenotype that
will be expressed in a kernel receiving the
set locus is predestined. The setting will
dictate the path of gene activity during
kernel development. Thus, the indirect
influence of the regulator, Spm, in control
of the setting process reflects still another
parameter of the mode of operation of a
gene-control system. It provides one
mechanism for inducing widely different
patterns of expression of a gene in differ-
ent parts of an individual organism.

The ‘impermanent settings just men-
tioned were made evident only in the
no-Spm kernel progeny of Spm-carrying
plants. It has been recognized for many
years, however, that nonheritable modifi-
cations affecting the action of a gene
under the control of a known system may
occur, and that they may be expressed in
cells that do carry the regulator of the
system. The component elements respon-
sible for these modifications were not so
readily analyzed as those involved in the
expression of states 7977B and 7995 of
a,"-2. Each such nonheritable modifica-
tion is known to arise as the consequence
of a single event occurring within a plant
cell. The cells derived from the mutant
cell carry the modification, whose effects
are registered as one type of altered gene
action in the progeny cells competent to
express this action. Germ cells arising
from the same initial cell, however, give
no indication that any change has
occurred; the modification appears to
have been erased in these germ-line cells.
In contrast to the preset patterns 1”
kernels, these nonheritable but obviously
preset types of gene control are expressed
in cells containing an active Spm element.
GENETICS RESEARCH UNIT

Examples of this kind of modification
were mentioned in Year Book 62. From
such observations it was possible to learn
that potential gene action can be set in
a cell early in development, so that many
cell generations later one particular grade
of gene action will be registered in only
one kind of tissue produced by the cell’s
descendants. Some of these nonheritable
settings result in a level of gene action
resembling that given by the wild-type
allele, whereas others effect reduced
levels.

The introduction to this report stated
that some of the known controlling ele-
ments in maize behave as independent
genetic agents, in that they contain
timing mechanisms for differential con-
trol of gene action during development.
It also stated that control systems are
extraordinarily efficient, a single system
having the potentiality to exert a wide
range of controls that affect not only
times but also types of gene action. The
preceding discussions, and those of earlier
reports, have attempted to illuminate the
many parameters of regulation of gene
action that may be ascribed to a single
system of controlling elements.

Cyclical Change in Phase of Activity
of Ae (Activator)

That the regulator element of the
Suppressor-mutator (Spm) system may
undergo changes in phase of activity was
first reported in Year Book 57. Only
recently was it learned that the regulator
element of the Activator (Ac) system
likewise may undergo changes in phase of
activity: from active to inactive and back
to active. This discovery was made during
an investigation of wz™-7, which is under
the control of the Ac system. Although
wx? had been isolated in 1952, from a
single kernel on an ear of a plant having
a,""* (Ac system) that was homozygous
for the wild-type Wx gene, only a rela-
tively few tests had been made with it
shortly after its isolation. Ae was known
to be located, initially, close to the locus

599

of wz™7. The purpose of the more recent
investigation, begun several years ago,
was to isolate instances in which Ac was
removed from this locus but the Wz gene
remained under the control of the Ac
system. In the course of the study, the
cyclically occurring changes in phase of
activity of Ac at the wz’ locus were
recognized.

The wz? locus was incorporated into
plants and kernels carrying various gene
markers located not only in the short arm
of chromosome 9, where wx-7 resides, but
also in other chromosomes of the comple-
ment. These combinations brought to-
gether in the nuclei of single kernels
different gene loci whose action is
controlled by the Ac system, with the
purpose of distinguishing between modifi-
cations at the wr’ locus that affect only
Ac and those that alter the expression of
the Wx gene, with or without an accom-
panying modification in the action of Ac
through change in its dose or state. A
change affecting only Ac would be detect-
ed by the like response of each of the
selected gene loci present in the nuclei of
a kernel. Modifications affecting the
expression of the Wx gene, accompanied
or unaccompanied by change in Ae, could
be identified, individually, by comparison
of the expressions of wx-7 with those of
the other Ac-controlled gene loci in the
kernel.

One of the combinations included a
particular state of a,"-°. Gene action at
the a,”°5 locus is under the control of the
Ac system. The selected state produces
uniformly light-pigmented plants and
kernels in the absence of Ac or when Ac
is present in an inactive phase. When Ac
is active, this locus gives rise in plants and
kernels to well defined areas that are
more deeply pigmented or are nonpig-
mented. Most of these areas consist of
descendants of cells in which a mutation
has occurred at the a,"-3 locus, and the
times and frequencies of such mutations
are known to be governed by the state
and dose of Ac present in the mutating
cells. After mutation the action of the
x

600

gene is no longer subject to control by the
Ac system.

Initial evidence of alternating cycles of
activity of Ac was seen in the kernel
progeny of a single plant in a culture of
sister plants that were a.™° Sho/a1 she;
Ac wz™7/wx in constitution. Each of
these plants had originated from a kernel
whose endosperm contained an active Ac
element to which both ai” and wr?
were responding. The plants in this
culture were crossed reciprocally with
plants that were homozygous for ai, she,
and wr and had no Ac. Some of them,
including the single one mentioned, were
also self-pollinated. This particular plant
was the only one in which Ac had entered
an inactive phase. The responsible event
must have occurred early in plant devel-
opment, because an inactive Ac was
present in the cells that gave rise to the
main stalk and tiller; and return to the
active phase occurred in only a few cells
of the plant.

The state of wz’ in this plant was one
that exhibits an intermediate level of Wx
gene action with inactive Ac, but higher
or lower levels when Ac is active. With
few exceptions, these changes in level of
gene action are not the consequence of
mutations at the wz™7 locus that result
in stable alleles. Subsequent changes
continue to occur as long as an active Ac
is present. This state also has allowed the
observation of a previously undetected
mode of control of Wx gene action during
endosperm development. The result is a
gradient of levels of Wz gene expression
in the mature endosperm, which is related
to the time of formation of cells during
development of the kernel. When kernels
are cut longitudinally and the exposed
cells stained with an I-KI solution,
differential staining of starch in the cells
of the endosperm can reveal different
levels of Wx gene action. Kernels with an
inactive Ac element have an inner core of
cells exhibiting a low level of Wz action ;
and from this core, toward the periphery
of the kernel, the level is seen to gradually
increase. In kernels having an active Ac,

CARNEGIE INSTITUTION

on the other hand, some cells within the
inner core evince a high level of Wz gene
action, and some cells nearer the periph-
ery disclose low levels. This mosaicism
interrupts the smoothness of the gradient
but does not obliterate it.

That Ac is inactive in the kernels
exhibiting an uninterrupted gradient of
Wax gene expression is shown by the
presence of pale pigment uniformly
distributed in the aleurone layer. In
particular cells of some kernels, Ac has
changed from an inactive to an active
phase, at times ranging from early to late
in the development of the kernel. The
descendants of such a cell form a well
defined sector, including part of the outer
layer of aleurone cells, in which Ac
activity is registered not only by modified
expressions of the Wx gene but also by
the response of a,"°*. A pattern of deeply
pigmented spots appears in the aleurone-
layer area included in the sector, the size
of the spots reflecting the time when Ac
became active. If it did so early in
development, the spots may be large; if
late, only small dots appear.

Investigation of changes in phase of Ac
activity at the wz? locus was continued
this year. One test was conducted to
examine the response of the wae? locus
having an inactive Ac to an active Ac
located elsewhere in the chromosome
complement. ‘Reciprocal crosses were
made between plants whose constitution
was a)" Sho/ai she; inactive-Ac w2r™7/
wa and plants that were homozygous for
ay, She, and wx and had one active Ac.
The a: locus responded to the intro-
duced active Ac in accordance with its
dose in the endosperm of the kernel. One
dose of Ac produced early-occurring
mutations at the locus; two doses, only
late-occurring mutations. The patterns of
response were the same in kernels having
we? as in those homozygous for the
standard wz allele. This observation made
+t evident that the inactive Ac at the
wa™7 locus was inactive not only with
regard to induction of mutation but also
with regard to contribution to dose
GENETICS RESEARCH UNIT

effects. The wr™’ locus also responded to
the introduced active Ac by producing
changes in Wz gene action, at times
during endosperm development that cor-
responded precisely to those of changes
at the a:"-* locus in the same kernel.
Evidence has not yet been obtained that
the times and frequencies of occurrence
of change in phase of Ac activity are
controlled by a mechanism similar to that
observed with Spm, but the preliminary
findings suggest it.

It must be stressed, in conclusion, that
our knowledge of genetic mechanisms
that control gene action during develop-
ment is still too limited to permit the

601

construction of reasonable models at the
molecular level to account for the wide
range in types of gene control described
here and in previous reports. The models
of gene regulation that have been pro-
posed to date are not yet fully defined,
and cannot be invoked in their present
form even though some of their aspects
are fitting. There are many parameters of
gene control during development. Aspects
of each must be clearly defined before an°
attack at the molecular level can be
initiated with hope of success. Awareness
of the problems at the genetic level is
prerequisite to their solution.

BIBLIOGRAPHY

Ledinko, N., Occurrence of 5-methyldeoxycyti-
dylate in the DNA of phage lambda, J. Mol.
Biol., 9, in press.

Mosig, G., Genetic recombination in bacterio-
phage T4 during replication of DNA frag-
ments, Cold Spring Harbor Symp. Quant. Biol.,
28, 35-42, 1963.