MUTABLE LOCI IN MAIZE

Barpara McCuintock

Previous reports have stated that a num-
ber of unstable loci have recently arisen in
the maize cultures. In a particular cell of a
plant, a normal “wild-type” locus becomes
altered; the normal, dominant expression
of this locus changes and gives rise to a
recessive expression (or, in several cases,
a recessive locus becomes unstable and
mutates toward a dominant expression).
This expression of the locus need not be
permanent. In some descendent cells, a
second change may occur within the locus
that results in the restoring of the capacity
of the locus to express the dominant pheno-
type or brings about an intermediate ex-
pression between full recessive and full

14

dominant. In the latter case, a third
alteration may occur in some descendent
cells that steps up the phenotypic expres-
sion toward the full dominant or reduces
it toward the full recessive.

When a locus changes from a stable to
an unstable state, recognition of the occur-
rence depends upon several factors. A
clear-cut phenotypic expression of the
changed locus is essential. This limits ease’
of recognition to those gene loci associated
with the production of some obvious plant
constituent, such as the chlorophyll pig-
ments or other plant pigments, or to con-
trasting morphological characters striking
enough to allow the contrasts to be de-
156

tected in single small sectors of some part
of the plant. The presence of a changed
locus that gives a recessive expression often
may remain hidden in a heterozygous
state until a self-pollination or some special
cross is made that allows its presence to
be uncovered. No attempt has been made
to detect the presence of these hidden
changed loci, and thus the frequency of
occurrence of such changes is not known.
The unstable loci that have been dis-
covered have all appeared in the progeny
of crosses designed for other purposes. It
is suspected that the number of mutable
loci still undetected may be quite great,
with numerous independent occurrences
of unstabilization of the same locus. The
detected unstable loci include many pre-
viously unknown to maize geneticists.
Some well known loci in maize have also
become unstable in these cultures. These
include two independently arising un-
stable yg loci (yellow-green chlorophyll),
four independently arising unstable ¢ loci
(c, colorless aleurone; C, colored aleu-
rone), three independently arising un-
stable wx (waxy) loci (wx, starch stain-
ing red with iodine; Wx, starch staining
blue with iodine), and one unstable a»
locus (anthocyanin in aleurone and plant).
With the exception of the two yg loci, all
these unstable loci have originated by a
change in a normal, previously quite stable
locus showing dominant expression. In
the case of unstable yg, the recessive
locus mutates to form chlorophyll of a
much darker color than the normal yg
locus expresses,

These mutable loci fall into two distinct
classes: (1) those that require the presence
of a second locus, the activator locus (Ac),
for instability to be expressed and main-
tained, and (2) those that do not require
such a second locus for instability to be
expressed. The Ac locus itself is unstable

CARNEGIE INSTITUTION OF WASHINGTON

and resembles in this respect the second
class of unstable gene loci.

Two recognizable subdivisions of the
mutation process are shown by all the un-
stable loci: (r) control of the time and the
apparent frequency of mutations in a
tissue by a factor capable of changing
during a mitosis so that altered rates of
mutation will be expressed in descendent
cells following such a change; and (2) the
subsequent change at the locus that occurs
in a particular cell and gives rise im-
mediately to recognizable altered pheno-
typic expression in this cell and its de-
scendent cells. The term “state of a locus”
has been used to distinguish the differing
potentialities of a mutable locus for ex-
pressing visible mutations in descendent
cells. The mutations giving rise to changes
in state of a locus are readily recognized
by the altered rates of the second type of
mutation that follow from such an event.
In the second class of mutable loci—those
controlled by A4c—the time and apparent
frequency of mutations of the affected
locus are controlled by the states of both
loci: the locus controlled by Ac and the Ac
locus itself.

During the past year, attention has been
directed mainly to those loci that require
the Ac locus for continued expression of
instability. All these loci that are known
are in the short arm of chromosome 9.
possibly because the genetic methods being
used allow certain mutable loci in this
arm to be readily detected. Those receiv-
ing particular study include the Ds (Dis-
sociation) locus, mentioned in previous
reports, two independently arising mu-
table ¢ loci, and two independently aris-
ing wx loci. In each case, the mutable
recessive state of the previously stable
dominant locus arose in a particular cell
of a plant that also possessed an Ac locus.
The origin of this potent Ac locus is quit
unknown. It has been traced to a culture
DEPARTMENT OF GENETICS

grown in the summer of 1942, but its
presence in still earlier cultures cannot be
traced. It cannot be stated that Ac induces
the initial state of instability of a locus,
although a causal connection may be
suspected.

None of these Ac-controlled mutable
loci will show any type of instability if
the Ac locus is absent from the nucleus.
As soon as Ac is introduced into a nucleus
having such a susceptible locus, instability
is resumed. Mutations at the susceptible
locus sometimes may occur in the initial
nucleus following the introduction of Ae.
To illustrate the control of mutability of
a locus by Ae, the mutable ¢” locus may
be used. This locus has been given the
symbol c”* because it was the first of the
several mutable ¢ loci isolated; the normal
c locus, used for many years in genetic
experiments, is designated c* because it is
a stable locus not mutating under the
influence of Ac. The behavior of this
unstable recessive c locus in the designated
constitutions is indicated below:

e’t ge: no mutations; aleurone layer color-

less

c? ac: no mutations; aleurone layer color-
less .

c* Ac: no mutations; aleurone layer color-
less

c™-1 4¢: mutations to C occur; aleurone layer
variegated for color

It should be emphasized that, with similar
constitutions, the same type of responses
will be registered by any of the other loci
that are activated by de.

Nature oF THE Ac AcTION

The Ds locus. The Ds locus was dis-
covered before any of the other Ac-con-
trolled unstable loci. Ds is located at the
position demarking the proximal third of
the short arm of chromosome g, between
one and two crossover units to the right

157

(toward the centromere) of the W= locus.
The normal, nonmutating locus has been
designated ds. An observable Ds mutation
arises as the consequence of breakage of
two sister chromatids within the Ds locus
in each case. This breakage is followed
by fusions of broken ends to give rise to
a U-shaped acentric fragment composed
of the two sister segments of the distal
two-thirds of the short arm, and a U-
shaped dicentric chromosome composed of
the proximal third of the short arm, the
centromere, and all of the long arm of
each of the two sister chromatids. Nearly
all the known loci of this chromosome are
carried in the distal two-thirds of the short
arm of chromosome g and are thus in-
cluded in the acentric fragment after a Ds
mutation. The genetic methods of ob-
serving the time and frequency of the Ds
mutations in various parts of the plant
have been explained in previous reports
(Year Books Nos. 45 [1945-1946] and 46
[1946-1947]).

The Ac locus acts upon Ds, or any other
locus susceptible to it, usually quite late
in the development of any of the sporo-
phytic tissues. In contrast, action of de
on susceptible loci may be apparent at all
stages in the development of the endo-
sperm tissues. In this respect, the endo-
sperm behaves like an extension of the
sporophytic tissues.

As stated above, the detectable Ds mu-
tations are associated with breaks at this
locus that are followed by 2-by-2 fusions
of those broken ends of sister chromatids
that lie adjacent to one another. This type
of fusion following breakage is not the
only one that could occur. Restitutions
could take place, re-establishing the pre-
vious organization of the chromatids; or
crisscross fusions of broken ends could
occur, simulating crossing over but in-
volving sister chromatids in somatic nuclei.
It is not known whether or not the latter
158

types of fusion do occur with expected
frequencies at the Ds locus. The evidence
suggests that they may occur, at least
occasionally. It is known that breaks at
the Ds locus may be followed directly or
indirectly by fusions of broken ends other
than those giving rise to the U-shaped
dicentric chromatids. This is shown by
the fusions that occur after an occasional
coincidental break in another part of the
chromosome. It is possible then for a
broken end associated with a Ds mutation
to fuse with one of these other newly
broken ends. This gives rise to a gross
chromosomal aberration that can be ana-
lyzed. Such chromosomal aberrations
have been observed in individual cells or
clusters of cells in the examined sporo-
cytes of the Ds Ac plants. Several of these
translocations have been found in .indi-
vidual plants, and have received further
study. These aberrations have been use-
ful in interpreting occasional incon-
sistencies in the behavior and the location
of the Ds and the Ac loci. It is now known
that the Ds locus may change its position
in the chromosome after such coincidental
breaks have occurred. One very clear case
has been analyzed, and, through appro-
priate selection of crossover chromatids,
strains having morphologically normal
chromosomes g have been obtained. As
a consequence of this aberration, the Ds
locus in these strains has been shifted from
a position a few units to the right of Wx
to a position between J and SA. This is
a very favorable position for showing the
nature of the Ds mutation process. When
the usual type of Ds mutation occurs at
the locus in this new position, a U-shaped
dicentric chromosome is formed. At the
succeeding anaphase a chromatin bridge
is produced, which: subsequently breaks.
The breakage-fusion-bridge cycle is thus
initiated and can be genetically expressed
in succeeding nuclear divisions. This is

CARNEGIE INSTITUTION OF WASHINGTON

because the loci of Sk, Bz, and Wx now
lie between Ds and the centromere. With
appropriate genic constitutions of chromo-
some 9, detection of the breakage-fusion-
bridge cycle is certain after a Ds mutation
has occurred in this new location. These
observations have been made, and the
genetic analysis completely confirms the
cytological analysis of the nature of the
Ds mutation process.

The inheritance of Ac as a separate
locus. Ac behaves in inheritance as a single
independent locus. Tests for the presence
of Ac in Fe progenies of ds ds Ac ac plants
have given ratios of 1 Ae Ac to 2 Ac ac to
1 ac ac; backcrosses of Ac-ac plants by
ac ac plants have given ratios of 1 Ac ac to
1 ac ac plant. Crosses of Ac Ac plants by
ac ac plants have given progenies all of
which were found to be Ac ac. In the
many crosses that have been made with
plants that were heterozygous for both
Ds and Ac, independent inheritance of
the Ac and Ds loci was clearly established,
except in three independently arising
cases. In these three cases, Ac was
obviously linked with Ds and could be
located 6 to 20 crossover units to the right
of it. Such linkages were maintained in
later tests of the progeny of one of these
plants, when both the crossovers and the
noncrossover chromatids were tested.
Similar tests are now under way for the
other two cases. Cytological examination
of heterozygous plants in each of the three
cases showed no observable chromosomal
abnormalities. Chromomere matching be-
tween synapsed homologues was perfect
for all of chromosome 9.

Either of two possibilities may explain
this change in genetic location of Ac.
First, 4c may be located toward the end
of the long arm of chromosome g and
normally show no genetic linkage with Ds
because the crossover distance is too great.
In the case of Ds-Ac linkage, a newly
DEPARTMENT OF GENETICS

arising crossover modifier, not associated
with a gross chromosomal abnormality,
would need to be invoked to explain the
observations. This modifier would need
to be closely linked with Ac, as it has not
been removed in the crossover tests so far
made. Second, the dc locus may have
been removed from its former position
and inserted into a new position in chro-
mosome 9 in a manner similar to that
observed for the transposition of the Ds
locus, described above. Because Ae in-
duces breaks at specific loci and gives
evidence of undergoing a specific breakage
process itself, this latter explanation is not
improbable. Tests are now under way
to determine whether or not either of
these alternatives can apply.

The effects of dosage of Ac. Studies of
the effects of various dosages of Ac have
shown that the time and apparent fre-
quency of Ac-controlled mutations is in
large measure a function of dosage of Ac.
This applies both to the sporophytic tissues
and to the endosperm tissues. The endo-
sperm tissue is 37. Here, one, two, or
three doses of Ac can be obtained and
observations ‘made of the effects of each
of these dosages on any of the Ac-con-
trolled mutable loci. All Ac-controlled
mutable loci respond alike to the various
Ac dosages. The description given below
of responses of the Ds locus in endosperm
tissues can be applied equally well to the
responses of any of the other Ac-controlled
mutable loci.

In Ds ds ds kernels, the mutation rates
of the single Ds locus have been compared
in Ae ac ac, Ac Ac ac, and Ac Ac Ac
endosperm constitutions. One dose of Ac
allows mutations to occur relatively early;
considerable irregularity in both time and
frequency of mutations is apparent, for
the endosperm is often divided into various
larger or smaller sectors each showing its
own special rate of mutation. This irregu-

159

larity is only apparent, as subsequent evi-
dence will show. ‘Two doses of de delay
the timing of Ds mutations so that they
occur relatively late in the development
of the endosperm. The frequency may be
very high, however. When three doses
of Ac are present, only relatively few,
very late-occurring Ds mutations are
recognized in these kernels. Three doses
of Ac appear to decrease the frequency of
Ds mutations, but this appearance is prob-
ably deceptive. It is more likely that the
increased dosage of Ac so delays the
timing of Ds mutations that only the
very earliest-occurring of these can pos-
sibly show in the endosperm, and that if
the endosperm cells had continued to
divide, very large numbers of Ds muta-
tions would have appeared. In other
words, the tissues mature before these
potential Ds mutations can be expressed.
The validity of this interpretation will be
apparent when the various states of the dc
locus are defined and their dosage re-
sponses compared. It may be concluded,

‘then, that the absolute dosage of Ac is one

main factor controlling the time and
apparent frequency of mutations: the
higher the Ac dosage, the later the occur-
rence of Ds mutations.

The frequency of Ds mutations is com-
plicated to analyze because of three factors:
(1) the maturing of tissues with high
dosages of Ac before Ds mutations can be
expressed; (2) changes in state of a single
Ac locus, resulting in changes in response
of the Ds locus that resemble either an
increase or a decrease in the dosage of Ac;
and (3) changes in state of the Ds locus,
resulting in changes in response to Ac
dosages. Factor (1) has been briefly con-
sidered above. Factor (2) will be con-
sidered next, and factor (3) will receive
brief consideration later.

Changes in state of the Ac locus. As
mentioned above, in the endosperms of all
160

kernels having the constitution Ds ds ds
and de ac ac there is great variability in
the time and frequency of Ds muta-
tions within a single kernel. Various num-
bers of relatively large or small sectors
may be present, each showing a special
time and frequency of Ds mutations, or
showing no Ds mutations at all. In most
of these kernels, a few relatively early Ds
mutations are observed. Several male
parents of the constitution Ds ds Ac Ae,
when crossed to ds ds ac ac female plants,
produced kernels with unusually favor-
able sectors of the above-mentioned types.
In over go per cent of the Ds-carrying
kernels, sectoring occurred early in the
development of the endosperm. Each
sector showed a very distinct type of muta-
tion pattern. The sectors in two of the
most common classes of kernel will be
described for the purpose of illustration.
In one class, the kernel was composed of
four very distinct sectors. In one sector,
no Ds mutations or only a few very late
mutations occurred; in the second sector,
the Ds mutations occurred late but rela-
tively frequently and were uniformly dis-
tributed within the sector; in the third
sector, the Ds mutations occurred at a
time midway in development of the endo-
sperm and at a medium rate; there were
fewer mutations per area than in the
second sector, and they occurred earlier.
The fourth sector arose from a nucleus
that had undergone a Ds mutation; no
further Ds mutations could be registered
in this sector. The second most frequent
type of kernel was composed of only three
sectors, similar to the first, second, and
fourth sectors of the above-mentioned
type. It had no sector of medium-timed
mutations. Nearly all the kernels resulting
from these crosses could be classified on
the basis of the presence of combinations
of two, three, or all four of these distinct
types of sectors. The origin of a sector,

CARNEGIE INSTITUTION OF WASHINGTON

in each case, could be traced to a nucleus
arising from the earliest divisions in the
endosperm. The ratio of kernels with
specific combinations of sectors was sur-
prisingly constant. Some combinations
occurred with high frequencies, others
with relatively low frequencies; and some
combinations were not found at all.

It is obvious that something occurs in
the early divisions of the endosperm that
results in nuclei with differing potentiali-
ties for expressing the time and frequency
of Ds mutations. Immediately upon ex-
amination of these kernels, one is im-
pressed with the resemblance of the muta-
tion patterns in the various sectors to the
patterns that have been obtained by com-
bining various dosages of Ac. Here, how-
ever, only one Ac locus was introduced
into the primary endosperm nucleus.
Obvious questions present themselves: Do
these sectors arise as the consequence of
some abnormality that results in segrega-
tions of various dosages of Ac in these
early nuclei? If so, what abnormalities
occur and what is the segregation mecha-
nism? Neither nondisjunction of chroma-
tids containing Ac, nor transposition of
the Ac locus to another position in the
chromosome set followed by mitotic seg-
regation, satisfactorily explains the ob-
served ratios of the various types of kernels.
A satisfactory fit is obtained if it is assumed
that (1) the Ac locus is composed of a
number of identical and probably linearly
arranged units, and (2) changes in the
number of units can take place at the locus
during or after chromosome reduplication
such that one chromatid gains units that
the sister chromatid loses. If the Ac-
carrying chromosome is double at the time
of entrance into the primary endosperm
nucleus, such an event involving one of
the two chromatids during the subsequent
division could result in four nuclei, two
with unaltered units within the Ac locus
DEPARTMENT OF GENETICS

and two with altered units within the Ac
locus (one of these having an increased
number of units and the other a decreased
number). The four resulting nuclei would
not be alike with respect to dosage of
units within the single Ac locus. If the
number of units within the Ac locus con-
trols the time and frequency of Ds muta-
tions in a manner comparable to known
dosages of whole Ac loci, the time and
frequency of Ds mutations in the various
sectors could be explained. If such a
change in units at an Ac locus is accom-
panied by a Ds mutation and involves
one of the four chromatids carrying Ds,
the kernels that have one sector produced
as the consequence of a Ds mutation could
be accounted for. This interpretation is
subject to several kinds of test. It should
be possible to predict the Ac type of con-
trol in those sectors that have undergone
an early Ds mutation. Special genetic
methods for detecting the Ae action in
these sectors have been devised. Prelimi-
nary results from one test indicate a good
fit with expectation; the most favorable
tests should be available in the summer’s
harvest.

Other evidence supports the assumption
that the Ae locus is composed of a num-
ber of identical units and that these units
can increase or decrease as the consequence
of a somatic mutation within the Ae locus.
This considers what has been called the
“state of the locus.” The various states
of the Ac locus are recognizable by the
time and frequency of the mutations it
produces when combined with one par-
ticular Ds locus. This particular Ds locus
must have a constant state. A Ds locus
can be maintained in a constant state by
keeping it isolated in an ac ac plant; for
without Ac the Ds locus is quite stable
and remains unchanged. dc loci have
been isolated that produce in Ac ae ac
constitutions relatively late Ds mutations.

161

When such late mutations result from a
single dose of Ac, the Ac locus is con-
sidered to be in a high state. Other de
loci have been isolated that in single doses
induce many early mutations. These are
called low-state Ac loci. In an Ae Ae ac
constitution (two doses) a low-state Ac
locus may give a mutation pattern re-
sembling that obtained from a high-state
Ac ac ac (one dose) constitution. Again,
the Ac Ac Ac (three doses) constitution
of a low-state Ac locus may give mutation
patterns resembling those produced by a
high-state Ac Ac ac (two doses) constitu-
tion. The effects produced by three doses
of a high-state Ac locus are most instruc-
tive, because the delay in timing of muta-
tions with this particular state and dosage
of Ac is so great as to result in no visible
mutations, or only an occasional one. The
tissues have ceased dividing before the
potential mutations can occur. The vari-
ously exhibited states of individually
selected Ac loci and the effects produced
by each in one, two, and three doses,
respectively, give added support to the
interpretation that the state of any one
Ac locus is an expression of the number
of reduplicate units present within this
locus. It is obvious that the Ac locus itself
is mutable, for Ac loci with different states
can be recognized and isolated in the
progeny of a cross of Ac ac by ae ac. It is
suspected that what can be seen in the
sectorial kernels described above are muta-
tions of the Ac locus.

Tue Murase ¢ Loci

Mutable c™*. One of the two mutable
c loci being studied originated from the
cross of a female plant homozygous for
c sh ds ac by a male plant homozygous for
C Sh Ds and heterozygous for Ac (Ae ac).
The same male plant was used in making
a number of similar crosses. On one of the
162

resulting ears, a single aberrant kernel was
observed. Instead of showing either a
complete C color (C Ds/e ds/c ds, ac ac ac
constitution) or a variegation pattern com-
posed of colored aleurone with colorless
sectors, owing to Ds mutations (C Ds/e
ds/c ds, Ac ac ac constitution), this kernel
was obviously composed of colorless
aleurone in which colored areas were pres-
ent. In pattern of variegation, it was the
reverse of expectation: a c locus appeared
to be mutating to C. This kernel was
removed, a plant was grown from it, and
various crosses were made to determine
the nature of the altered expression of
aleurone color. Appropriate genetic anal-
ysis indicated that a mutable ¢ locus was
present which had arisen from the normal
C locus present in the male parent. This
C locus had changed to a mutable ¢ locus
capable of mutating back to the original
C. This mutable ¢ locus is Ac-controlled
and responds in precisely the same manner
as the Ds locus to various doses of Ac and
to various states and observable changes in
state of the Ac locus. With respect to Ac,
the responses of the two loci are amazingly
alike. Ds mutations, however, are known
to be the result of some mechanism lead-
ing to chromosomal breakage and fusion,
whereas the mutations at the ¢ locus appear
to involve quite a different mechanism,
leading to changes in expression of the
locus from recessive to dominant.

Similarities in response of Ds and c™}
to Ac, together with the known breakage
mechanism at the Ds locus and also the
obvious changes in state of the Ac locus
that can be explained on the basis of a
chromosomal breakage mechanism, lead
one to suspect that some kind of chro-
mosomal breakage and fusion mechanism
may likewise be .responsible for these
reverse mutations.

Because ¢”* and ¢”? are very different

CARNEGIE INSTITUTION OF WASHINGTON

in their phenotypic expressions, however
similar they are in other respects, it should
be stressed that visible mutations at the
e™* locus always result in the expression
of a full dominant phenotype. Varying
expressions of color intensity to either side
of that expected from a single dose of
normal C do not appear.

Mutable c™?. The second mutable ¢
locus, ¢”~, arose in a somatic cell of a plant
of the constitution 1 Sh ds/C Sh ds, Ac Ac.
The plant was sectorial for this mutable
¢ locus. With respect to dosage of Ac, the
production of sections, etc., the ¢”-* locus
responds exactly as do the Ds and c™?
loci. The phenotypic expression of muta-
tions at the ¢””* and at the c”? locus is
distinctly different. The mutations of
c”? give rise to sectors showing great
variation in color intensity, from a very
faint to an intensely deep color—often
much deeper than that produced by a
single, double, or even triple dose of C.
Any one intensity may appear as a mutant
sector in any part of the aleurone layer
on any one kernel; also, any one kernel
may show a number of different mutations
each having its own particular color
intensity.

The very great differences in phenotypic
expression of the two independently arising
mutable ¢ loci suggest that the normal
C locus may be composed of at least two
blocks, each having its own particular
organization and function but both neces-
sary for producing the C phenotype. They
are not strictly complementary, for com-
binations of c™* and c”™? in the same
endosperm do not result in the production
of C color. The mutable ¢”* locus is
assumed to have arisen from a normal C
locus after alterations had occurred in only
one block. On the other hand, the mutable
c™* arose from a normal C locus after
alterations that involved only the second
block. The ¢”* locus undergoes muta-
DEPARTMENT OF GENETICS

tional changes only in the one affected
block, whereas the mutations of ¢”” in-
yolve changes only in the second block.
A possible alternative to this interpreta-
tion will be considered in the concluding
paragraphs of this report.

Tue Murasre wx Loct

Two mutable wx loci are now under
investigation. Each was found as a single
aberrant kernel on an ear resulting from
the cross of a plant homozygous for wx
ds and ac by a plant that was homozygous
for Wx and also for Ac. A different male
parent was involved in each cross. The
first case found, wx", was recognized
because the aberrant kernel showed a wx
phenotype with Wx sectors and spots
instead of the expected full Wx phenotype.
Because other well marked genetic char-
acters were involved in this cross, it could
be determined that no contamination had
occurred to give rise to this kernel and
that the mutable wx condition had arisen
at the Wx locus in the chromosome 9 con-
tributed by the male parent. No known
mutable wx loci were present in any of the
stocks at the time of discovery of the
original kernels of either wx™* or wx”.

The second case, wx”, was first. ob-
served as an aberrant kernel because it
showed a recessive wx phenotype, instead

of the expected Wx. No mutability was °

recognized in the original kernel. The
mutability of the locus was recognized
later, when the pollen of the plant arising
from this kernel was examined.

A third case of a mutable wx locus
arising from a normal dominant Wx locus
was observed in the cross of a female
Wx Wx, Ac Ac plant by a male wx wx,
ac ac plant. The kernel was almost com-
pletely wx, with only small specks of Wx.
The mutations to Wx occurred late in the
development of the endosperm. This type

163

of mutation pattern suggested that the.
mutability was Ae-controlled; for two
doses of Ae were present in the endosperm,
and two such doses are known to give
this pattern with all Ac-controlled mutable
loci. The embryo in the kernel did not
germinate, and so this case of an inde-
pendently arising mutable wx locus has
been lost.

The two recovered mutable «x loci have
received only preliminary study because
of their recent origin. The wx" + locus is
‘Ac-controlled and responds to Ac in all its
various aspects as do the Ds, o™, and e™?
loci. It is not known whether wx” is
controlled by Ac.

All Ac-controlled mutable loci show
mutations only late in the development of
the sporophytic tissues. In the sporogenous
cells, such mutations may be delayed until
the premeiotic or meiotic nuclei are
formed. The starch in the pollen of plants
homozygous for a normal wx locus stains
a clear red-brown color with dilute solu-
tions of iodine and potassium iodide. The

starch grains in the pollen of plants homo-

zygous for a normal Wx locus stain a dark
blue color with this same solution. In
heterozygous plants, as expected, two
types of pollen grains are present in equal
proportions, those staining red-brown (the
wxearrying grains) and those staining
deep blue (the Weearrying grains). In
the two original plants carrying these
mutable wx loci,a normal nonmutating
wx locus was present in one chromosome
g and the mutable wx locus in the homol-
ogous chromosome 9. Examination of
the pollen of these two plants gave the
first indication that the primary mecha-
nism responsible for mutations of the. wx
loci could be associated, at least in some
cases, with chromosome breakage and
fusion. Ih both plants, most of the pollen
grains, when stained with iodine, were
visually wx (red-brown) in phenotype.
164

In the plant carrying wx", however,
about 2 to 3 per cent of the grains were
clearly different in staining reaction. These
stained blue with iodine, the intensities
of color ranging from light gray-blue to
the very deep blue characteristic of pollen
grains carrying a normal Wx locus. A
high percentage of these blue-staining
grains, often more than 50 per cent, were
small and only partially filled with starch
granules. This contrasted greatly with
the proportion of such defective grains
(only 2 to 3 per cent) in the wx-staining
class. The pollen of the plant carrying
wx"? differed from that of the plant
carrying wx™* in that the recognizable
blue-staining grains were confined mainly
to the defective class. Less than 1 per cent
of the pollen grains in this plant were
defective, and in approximately half of
these the starch granules stained blue.
Here, also, the intensities of color ranged
from a light gray-blue to a very deep blue.
Only a very few of the normal-appearing
pollen grains showed any trace of blue
color, and it was always very faint. This
contrasted greatly with the pollen of the
wx™™” plant, where many. of the deeply
blue-staining grains were normal in ap-
pearance. The questions arise: (1) Why
is the proportion of defective grains so
high in the mutated class in both plants?
(2) What is the reason for the various
intensities of the staining reaction? (3)
Why are the normally developed grains
with mutated loci so different in staining
reaction in these two plants?

An interpretation, subject to further
testing, has been formulated. It assumes
that the normal Wx locus is composed of
a number of identical, reduplicated units,
each unit contributing its share in the
ultimate conversion of precursor starch to
the amylose type characteristically pro-
duced by the Wx locus. The more units
are present at the locus, the greater is the

CARNEGIE INSTITUTION OF WASHINGTON

conversion and the stronger the iodine
staining reaction. Published quantitative
chemical studies (Sprague, Brimhall, and
Hixon, Jour. Amer. Soc. Agric., vol. 35,
pp. 817-822, 1943; Brimhall, Sprague, and
Sass, Jour. Amer. Soc. Agric., vol. 37, pp.
937-944, 1945) of the nature of the starch
in endosperms having various genic con-
stitutions of wx, Wx, and a low allele of
Wx known as wx? have given a sound
basis for assuming a quantitative reaction
of the Wx locus. The starch in wx wx wx
endosperm is composed of amylopectin.
With one dose of Wx (Wx wx wx con-
stitution), about 18 per cent of the starch
is synthesized to amylose, the remainder
being amylopectin. With two and three
doses of Wx, more starch is synthesized
to amylose, but not in proportion to
dosage (20 and 22 per cent, respectively).
Tests of wx*, however, have shown a pro-
portionality with dosage. Here, one dose
of wx* (wx* wx wx constitution) gives
0.65 per cent amylose, and two and three
doses give 1.20 and 2.40 per cent, respec-
tively. These published results suggest
that the number of reduplicated units in
the normal Wx locus may be high enough
to approach, in a single dose, an effective
utilization of the available substrate to be
converted. This is apparently not so for a
single dose of the wx allele. Here, too
few units may be present and much of
the available substrate may be unutilized.
This available substrate could be utilized
if more gene units were present. Added
doses of the wx* allele could accomplish
this. On the assumption that different
numbers of gene units are present in the
Wx and wx" alleles, it is possible to explain
the different dosage responses of these two
alleles.

According to the hypothesis that the
normal Wx locus is composed of a num-
ber of reduplicated units, it is assumed
that the mutable wx loci arose from a
DEPARTMENT OF GENETICS

normal Wx locus by loss of units from the
locus. The remaining units are too few
to produce a visual change in the color of
the starch when stained with iodine. If
the mutation process results in an increase
of units within the locus, enough amylose
may be produced to give a visible color
reaction with iodine. Since the wx” locus
gives a light orchid stain in the endosperm
and a lighter stain than full Wx in the
pollen grain, the number of residual units
in the two mutable wx loci should be even
lower than that present in wx*, Because
sndividual mutations occurring at the
we? and wx"? loci are not alike, but
result in each case in the production of
one particular intensity of color reaction,
‘t must be assumed that the increase in
the number of units at any one mutation
is not constant; some mutations result in
only small increases, others in larger in-
creases in number of units, and so on. The
various grades of intensity of staining re-
flect these variable increases in the num-
bers of units: the more units are present,

the more amylose is produced, and con- .

sequently the more intense is the staining
reaction.

It is believed that the defective grains
‘n the mutated class could arise as the
consequence of a breakage-fusion mecha-
nism that results in U-shaped dicentric
chromatids and an acentric U-shaped frag-
ment, as described for the Ds mutations.
If this break occurred within the wx locus
or to the left of it, the dicentric chromatid
could have more amylose-producing units
in it than either one of the original chro-
matids: increased or doubled in the first
case, and doubled in the second case.
Breakage of this bridge in the succeeding
anaphase could occur nonmedially and
give rise to one broken chromatid having
all these increased numbers of units and
a sister chromatid having none.’ Suc-

165

cessive breakage-fusion-bridge cycles could
double, triple, quadruple, etc., the number
of units. Only a few mitoses would be
required for building up large increments
of such units. Are the defective Wx-
staining grains the result of this mecha-
nism? Are they defective because they
have lost segments of the short arm of
chromosome g that always accompany the
formation of such dicentric chromatids?
Will a deficiency of one-half to two-thirds
of the short arm of chromosome 9 give
rise to a detectable pollen grain, and will
it be of this particular defective type?

An affirmative answer to this last ques-
tion has been obtained by examining the
pollen of Wx Ds/Wx Ds, Ac Ac plants.
This pollen should all be Wx-staining
except for the grains produced after a
premeiotic or meiotic Ds mutation; these
should be deficient for at least two-thirds
of the short arm of chromosome 9- If, in
these grains, some starch granules are
formed, the starch should give the wx
staining reaction, because a Ds mutation
would place the Wx locus in the acentric
fragment. This fragment is usually lost
to the nucleus in the division following
the Ds mutation. The dicentric chromo-
somes in the nuclei formed after a Ds
mutation should have no Wx locus at
all. It has previously been determined,
through other types of experiments, that
a deficiency of the Wx locus will give rise
to starch of the wx staining type. By
examination of the pollen of Wx Ds/Wx
Ds, Ac Ac or Ac ac plants, it has now been
determined that such small, partially filled,
wx-staining grains are present in the
relatively high numbers expected. Only
rarely does one find such grains in the
pollen of Wx Wx ac ac plants, and those
found may arise from occasional adjacent
fusions of broken ends after crossover
breakage at meiosis. There is no obstacle,
166

then, to considering that the defective Wx-
staining grains in the pollen of the wx”
and wx plants have arisen as the conse-
quence of dicentric chromatid formation
followed by the breakage-fusion-bridge
cycle. The Wx locus is to the left of Ds,and
the segment of chromosome g absent from
these grains is even smaller than that
found after a Ds mutation. The various
intensities of the Wx staining reaction in
the defective class of pollen grains may
merely represent the various dosages of
Wx units present in these grains; for
various dosages could be anticipated.
The breakage-fusion-bridge cycle alone
cannot explain many of the mutations
from wx to or toward.a Wx phenotype.
Normal, functional Wsx-staining pollen
grains are present in the wx” plant. It
is possible that these are the reciprocal
products of a breakage mechanism’ that
sometimes gives rise to dicentric chro-
matids. A crisscross type of fusion may
occur between broken ends after breakage
at one locus in two sister chromatids. If
the breakage occurs within the compound
locus at unequal positions in the two
chromatids, one resulting chromatid might
gain enough units at a single breakage
and fusion to produce a quantity of amy-
lose starch that would give a detectable
staining reaction. Several successive mu-
tations could step up the unit dosage with-
in the locus considerably. A graded series
of staining intensities could be expected
to follow such events. Examination of
the mutation process in the endosperm
tissues carrying c”* and wx" suggests
that such successive mutations may occur.
Examination of the kernels result-
ing from crosses involving the wx"?
locus has been illuminating. It has
shown: (1) that mutations of wx"
occur, as expected, in the endosperm
tissues; (2) that these mutations differ in
phenotype, some giving rise to sectors

CARNEGIE INSTITUTION OF WASHINGTON

showing only very faintly staining starch,
others to sectors showing intensely blue-
staining starch, with all others falling be-
tween these extremes; (3) that at least
five grades of intensity can be recognized
among the mutations within a single
kernel; and (4) that twin sectors fre-
quently occur, one sector showing a deeper
blue stain than the sister sector. Exami-
nation of these kernels also showed that
a single mutation may be followed in
successive divisions by further mutations,
and that these latter result in starch with
either an increased or a decreased staining
intensity. In other words, the mutation
process may result in changes in either
direction: toward full dominant or toward
full recessive. A few sectors arising as the
consequence of the breakage-fusion-bridge
cycle were present in these kernels. Large
areas within these sectors gave the wx
staining reaction, but there were subsectors
showing the Wx staining reaction. With-
in these subsectors, a wide variation in
intensity of staining reaction was ob-
served, The spatial relations of the cells
showing these varied intensities of stain-
ing reaction were instructive; for adjacent
areas probably arising from sister cells and
showing inverse relations of color intensi-
ties were most obvious. The differences
in intensities of staining reaction in these
twin areas ranged from slightly detectable
to pronounced, with extremes showing a
deep blue adjacent to a red-brown (wx).
This latter observation strongly supports
the interpretation of the relation of dosage
to the intensity of the staining reaction.
Dosage changes appear to be responsible
for the changes in staining reaction within
these sectors having dicentric chromatids;
the greater the dose, the more intense the
staining reaction, and vice versa. It is
realized that increments or decrements of
units must fall within the range that can
DEPARTMENT OF GENETICS

give visually distinguishable changes in
intensities of color.

It is not mere speculation, therefore, to
consider that the quantitative grades of
mutation occurring at the mutable wx
loci may follow increases in the number
of identical units within a depleted locus
that is normally composed of a relatively
jarge number of such reduplicated units.
The question arises, What is the primary
mechanism responsible for such increases
and decreases, and do all mutable loci
reflect this same general mechanism?

CoNCLUSIONS

It may be premature to consider in
detail the question asked in the concluding
sentence above. Because so many mutable
loci have recently appeared in the maize
cultures, because in many respects they all
behave in very much the same way, and
because this behavior is similar also to
that described for other mutable loci both
in maize and in other organisms, it may
be profitable to review briefly the perti-
nent facts about the cases described in this
report, in order to ascertain the similari-
ties and dissimilarities among these cases.
They involve the mutable locus Ac and
the mutable loci it controls—Ds, emt, o™,
and wx".

Ds stands alone in that the chromosomal
consequences of a mutation at this locus
are known. The detectable Ds mutations
unquestionably arise as a consequence of
some mechanism that either brings about
breakage and fusion between sister chro-
matids at the Ds locus or simulates this
mechanism in its consequences; for dicen-
tric chromatids are produced after a Ds
mutation. In the pattern of mutations, om
and Ds are strictly comparable. Variega-
tion patterns produced by Ds mutations in
C ds/C ds/I Ds, Ac ac ae constitutions

can be so similar to variegation patterns

167

produced by ¢ to C mutations in ct",
‘Ac ac ac constitutions that they may be
indistinguishable by mere observation of
the kernels. Yet the formation of dicentric
chromatids alone cannot explain most of
the observed c”? mutations. This like-
wise applies to the ¢™” mutations and to
many of the mutations of wx™*, Some
form of chromosome breakage and fusion,
however, may be involved.

As stated earlier, mutations of c™? and
em? differ greatly in phenotypic expres-
sion. Visible mutations of ¢™* give rise
each time to the full dominant expression
expected from a single dose of C, whereas
those of c™-? are quantitatively expressed,
with color intensities varying from faint
to extremely deep. The normal C locus
is known to give dosage effects: the more
whole C loci present, the greater the depth
of color. By means of duplications of the
short arm of chromosome 9, it has been
possible to observe effects of doses up to
and including six C loci. With the highest
dose, the color of the aleurone is unusually

‘deep. Some of the mutations of c”? result

in intensities greater than that shown with
three doses of the normal C locus, whereas
others are so faint that they obviously have
produced much less pigment than is pro-
duced by a single normal C locus. The
various consequences of mutations of the
c? locus are in complete agreement with
the hypothesis that the mutations result
from graded increases in the number of
units within a depleted locus and that they
express themselves phenotypically by
graded increases in the substance or sub-
stances responsible for the phenotypic char-
acter. The observed mutations of c™* do
not lead to this hypothesis, for they show
no quantitative subdivisions. Similarity
of the c mutations to Ds mutations in
basic response to Ac, and their conformity
to the general pattern of the Ac-controlled
168

mutations that do give quantitative effects,
has led to formulation of a subsidiary
hypothesis rather than abandonment of
the general hypothesis. The validity of
this subsidiary hypothesis is subject to
tests that are now being conducted. As-
suming the fundamental mutation process
to be similar for all Ac-controlled mutable
loci, the genes in the mutating ¢”* block
of the C locus could be related to some
chemical process that requires a threshold
number of units for expression to be ful-
filled; or it may be that the mechanism
bringing about a change in units at this
mutable ¢ locus assures a specific increase
in number of units.

The mutations occurring at the c”-? and
wx" Joci are amazingly similar in every
respect. The mutations fall into a graded
quantitative series in both cases. Also,
both give twin or adjacent sectors showing
the same grades of contrasting intensities
of expression of the dominant character.
There can be little doubt that the same
mechanism is responsible for the muta-
tions occurring at these two different loci.
With regard to the wx" locus, there is
evidence for believing that the unit num-
ber within the locus may be responsible
for the expression of unit increases of the
dominant phenotypic expression. Con-
sidering the obvious similarities between
the two cases, it would be difficult to avoid
concluding that the same conditions apply
to mutations at the c”* locus. In this
connection it may be recalled that muta-
tions at the Ae locus likewise suggest a
mutation mechanism involving changes
in the numbers of units at the locus.

Because many of the mutations occur-
ring at the two mutable ¢ loci, the wx"?
locus, and the mutable Ac locus do not
result in dicentric chromatids, as do Ds
mutations, and do not lead to detectable
gross chromosomal aberrations, the mecha-
nism, if it is a breakage phenomenon,

CARNEGIE INSTITUTION OF WASHINGTON

must restore the normal chromosome
morphology. Unequal breakage within the
locus, followed by the crisscross type of
fusion mechanism, could accomplish this
end.

Any mechanism that gives rise to an
increase in numbers of units at a locus
might also give rise to the reverse con-
dition; that is, to chromatids with de-
creased numbers of units. If so, chromo-
somes with loci having various unit num-
bers should appear as isolates in these
cultures. Some of these isolates should
show more visible mutations than others
under given conditions. The frequency of
visible mutations would depend upon the
initial number of units present in the
locus before mutation occurred. The more
units were present, the greater would be
the chance that the increase in the units
during any one mutation would be suff-
cient to produce a visible effect; and the
converse would also be true. Isolates from
the different mutable loci, showing just
these expected variations in the rates of
visible mutations, have been obtained. The
term “state of the locus” applies as well
to the mutable loci controlled by Ac as to
the Ac locus itself. What has been termed
a high-state locus gives high rates of muta-
tion, and a low-state locus gives low rates
of mutation. It is possible, therefore, that
the number of units present at a mutable
locus is correlated with the state of the
locus as well as with the expression of
visible effects. It is a matter of degree. If
the initial number is high, but not high
enough to produce a visible effect, the
state of the locus may be considered high.
Conversely, if only a few units remain in
the locus, the state of the locus may be
considered low.

The above conclusions are supported by
the many observations that have been
made especially of the chlorophyll-produc-
ing types of mutable loci. As mentioned in
DEPARTMENT OF GENETICS

previous reports, the motivation for this
study of mutable loci was the observation
of twin sectors, apparently arising from
sister nuclei, that showed inverse rates of
visible mutations. The most extreme of
these twin sectors showed a mutation to
dominant in one sector, and in the sister
sector either a complete recessive or a very
much reduced rate of mutation as com-
pared with that in the surrounding tissue.
In these observed cases of twin sectoring,
it is obvious that the factor or factors con-
trolling the rate of mutation and the
visible mutations themselves are of the
same general nature, if not actually differ-
ent resultants of the same mechanism. The
factor responsible for twin sectoring acts
at a mitosis, and the apparent result is
that one chromatid gains something that
the sister chromatid loses. In the inter-
pretation given, this gain and loss are
considered to be an increase and a de-
crease, respectively, of identical units at
the locus, which, in turn, controls not
only the appearance of visible mutations

but also the state of the locus as reflected .

in the time and frequency of occurrence
of future visible mutations.

The above interpretations are being
used as a working hypothesis in continu-
ing studies of mutable loci. The evidence
at present points toward the presence of
reduplicated units within a locus, these
units often expressing themselves in a
quantitative manner through unitary ac-
tion on substrates responsible for pheno-
typic characters. The hypothesis that the

169

phenotypic expression of a locus depends
upon the number of such units present
in any one chromosome, and that some
mechanism or mechanisms can alter this
number and give rise to visible mutations,
is both sufficiently simple and sufficiently
integrative to afford precise direction in
some types of. experimentation. It is be-
lieved that this approach to one phase of
the over-all mutation problem may be
productive, even though it is realized that
the details will need clarification and may
be subject to degrees of modification. With
so many mutable loci behaving in very
much the same manner, it is unlikely that
many different, unrelated mechanisms are
involved. It is more likely that one general
type of condition exists in all these mutable
loci and that this condition may be altered
in any of these loci by one kind of
mechanism.

In this report, it has not been possible
to include a discussion of the many other
observations and conclusions that are
relevant to the subject, such as the restabi-
lization of a mutable locus, the behavior
of Ac-controlled mutable loci when two
or more are present in the same nucleus,
or the changes at the Ac locus that often
appear to accompany an Ac-induced muta-
tion. Nor has it been possible to describe
the accumulated evidence on some of the
non-Ac-controlled mutable loci, or to men-
tion the many gew mutations that are
constantly arising in these cultures. These
studies are beifig continued and will be
reported later,