-
Copyright 0 1994 by the Genetics Society of America
Modification of the Drosophila Heterochromatic Mutation
brownDominant by Linkage Alterations
Paul B. Talbert*91, Cosette D. S . LeCiel*'* and Steven
HenikofP".? *Fred Hutchinson Cancer Research Center and THoward
Hughes Medical Institute, Seattle, Washington 98104
Manuscript received August 13, 1993 Accepted for publication
October 8, 1993
ABSTRACT The variegating mutation brownDmniMnf (bp) of
Drosophila melanogaster is associated with an insertion
of heterochromatin into chromosome arm 2R at 59E, the site of
the bw gene. Mutagenesis produced 150 dominant suppressors of bwD
variegation. These fall into two classes: unlinked suppressors,
which also suppress other variegating mutations; and linked
chromosome rearrangements, which suppress only b p . Some
rearrangements are broken at 59E, and so might directly interfere
with variegation caused by the heterochromatic insertion at that
site. However, most rearrangements are translocations broken
proximal to bw within the 52D-57D region of 2R. Translocation
breakpoints on the X chromosome are scattered throughout the X
euchromatin, while those on chromosome 3 are confined to the tips.
This suggests that a special property of the X chromosome
suppresses b d variegation, as does a distal autosomal location.
Conversely, two enhancers of bp are caused by translocations from
the same part of 2R to proximal heterochromatin, bringing the b d
heterochromatic insertion close to the chromocenter with which it
strongly associates. These results support the notion that hetero-
chromatin formation at a genetic locus depends on its location
within the nucleus.
C HROMOSOMES of Drosophila melanogaster and many other
eukaryotes are composed of two distinct kinds of chromatin: the
pericentric regions consist of heterochromatin, which remains
condensed throughout the cell cycle, and the chromosome arms
consist of euchromatin, which decondenses during interphase. A
large body of work has shown that DNA sequences in heterochromatin
are largely repetitive and contain relatively few identifiable
genes; the bulk of functional genes reside in euchromatin (for a
re- view, see GATTI and PIMPINELLI 1992). In Drosophila embryos,
centromeric regions coalesce at the apex of blastoderm nuclei, and
heterochromatinization of these regions appears to occur around the
time of cellularization (RABINOWITZ 1941). In salivary gland
nuclei, the pericentric sequences are underrepre- sented relative
to the polytenized euchromatic arms and the coalesced centromeric
regions form a heter- ochromatic chromocenter (RUDKIN 1969; LAIRD
et al. 1973). The functional significance of this nuclear
compartmentalization of heterochromatin is not well understood.
Evidence for incompatibility of euchromatin and heterochromatin
comes from chromosome re- arrangements that juxtapose the two. Such
chromo- some rearrangements typically exhibit a variable cis-
inactivation of adjacent euchromatic genes, known as
15, Seattle, Washington 98195.
50, Seattle, Washington 98195.
Genetics 136: 559-571 (February, 1994)
' Present address: University of Washington, Department of
Botany KB- * Present address: University of Washington, Department
of Genetics SK-
position-effect variegation (PEV). For example, when a
rearrangement places heterochromatin next to the wild-type white
(w+) gene, which is necessary for pig- mentation of the eye, the
inactivation can be visualized as a variable pattern of pigmented
and nonpigmented ommatidia in the eye (for reviews, see SPOFFORD
1976; HENIKOFF 1990; SPRADLING and KARPEN 1990; REU- TER and
SPIERER 1992). Studies of salivary gland chromosomes show that
cis-inactivation of euchro- matic genes correlates with a change in
cytological appearance from euchromatic to heterochromatic
(HARTMANN-GOLDSTEIN 1966; HENIKOFF 198 1 ; HAY- ASHI et al. 1990;
KARPEN and SPRADLING 1990; UM- BETOVA et al. 1991). This has been
interpreted as a spreading of heterochromatin formation across the
rearrangement breakpoint into normally euchromatic adjacent genes,
causing their inactivation. Since PEV results from gene
inactivation, PEV alleles are nor- mally recessive to their
wild-type homologs.
PEV can be suppressed by dominant mutations in a large number of
euchromatic loci, as well as by an extra Y chromosome (reviewed by
EISSENBERG 1989; GRIGLIATTI 1991; REUTER and SPIERER 1992). Many of
the suppressor loci, known as Su(uar)s, show dosage effects such
that a single dose of a Su(var)+ gene will suppress PEV relative to
the normal two doses; a third dose will enhance PEV. Conversely,
some dominant E(uar)s enhance when wild type in one dose and sup-
press when in three doses. T o account for these ob- servations, a
mass-action model has been proposed in which dosage-sensitive
suppressor loci encode proteins
-
560 P. B. Talbert, C. D. S. LeCiel and S. Henikoff
that complex with DNA to form heterochromatin (LOCKE et al.
1988). Conversely, enhancer loci might encode proteins involved in
limiting heterochromatin formation. The process of heterochromatin
formation is then dependent on the concentrations of each of these
component proteins. Other Su(var) and E(var) genes that do not show
dose effects may encode en- zymatic functions that are necessary to
assemble chro- matin.
A form of PEV complementary to the cis-inactiva- tion of
euchromatic genes occurs for genes that nor- mally reside in
heterochromatin (BAKER 1968). The light ( I t ) gene at the base of
chromosome arm 2L apparently requires a heterochromatic environment
to function since it variegates when rearrangements move it to
distal euchromatin (HESSLER 1958; HILLI- KER and HOLM 1975;
WAKIMOTO and HEARN 1990). Interestingly, some suppressors of PEV of
euchro- matic genes act as enhancers of It variegation, as would be
expected if they limit heterochromatin formation (HEARN et al.
1991). The fact that no variegating rearrangements have been
observed that move It to proximal euchromatin has led to
suggestions that proximity to a centromeric compartment in the nu-
cleus may also be an important determinant of heter- ochromatin
formation (WAKIMOTO and HEARN 1990). A study of the heterochromatic
rolled ( r l ) gene has shown that it variegates when it is present
in a small block of heterochromatin surrounded by euchroma- tin,
but regains more normal function when it is moved near a large
block of heterochromatin, regard- less of its distance from the
centromere per se (EBERL et al. 1993). These studies argue that
distance from other heterochromatic elements is an important de-
terminant of PEV for heterochromatic genes (BAKER 1968).
In contrast to the situation for PEV of heterochro- matic genes,
involvement of nuclear localization in PEV of euchromatic genes is
less clear. MULLER (cited by EPHRUSSI and SUTTON 1944) was the
first to predict that differences in chromosomal positioning in so-
matic cells might account for PEV of euchromatic genes. Indeed,
numerous early studies showed that PEV mutations can often be
reverted by re- arrangements in the heterochromatin adjoining a
var- iegating gene, usually relocating that gene to a eu- chromatic
site (DUBININ 1936; PANSHIN 1938; GRIF- FEN and STONE 1940;
KAUFMANN 1942; HINTON and GOODSMITH 1950). Molecular studies with
the chro- mosomal inversion Zn(l)w"', which moves the w+ gene
proximally next to a disrupted block of pericentric
heterochromatin, have indicated that reinversions placing thew+
gene back to a distal location revert the variegating phenotype,
even though some heterochro- matic repeat sequences remain adjacent
to the w+ gene (TARTOF et al. 1984). Similarly, in reversions
of
T(1;2)dorUar7, at least 20 kb of heterochromatin re- mains
adjacent to the deep orange gene (POKHOLKOVA et al. 1993). In all
of these cases, however, it is difficult to evaluate whether the
reversions resulted from the change in position relative to a
chromocentral com- partment, the change in the amount or kind of
adja- cent heterochromatin, or all of these. Only DUBININ (1936)
reported revertant chromosomes in which a block of heterochromatin
and the adjacent breakpoint were moved to new locations by
euchromatic re- arrangements that left the heterochromatic block
in- tact.
Here we provide evidence that altering the chro- mosomal
location of euchromatic genes subject to PEV can have striking
phenotypic consequences, sim- ilar to what is seen for PEV of
heterochromatic genes. These findings were made as a result of
attempts to obtain dominant suppressors of PEV of the brown (bw)
gene, which controls production of the pteridine pigments of the
eye, using the variegating allele
being a spontaneous insertion of pericentric hetero- chromatin,
rather than an X-ray induced chromosome rearrangement (SLATIS
1955). Insertion of hetero- chromatin occurred distally in
chromosome arm 2R, which otherwise retains its normal gene order.
There- fore, euchromatic rearrangements of 2R can change the
position of b d relative to its centromere without simultaneously
altering the block of heterochromatin that induces PEV.
Variegating bw alleles not only inactivate the bw gene on the
rearranged chromosome (cis-inactiva- tion), they also inactivate
the bw gene on the homol- ogous chromosome (trans-inactivation;
reviewed by HENIKOFF et al. 1993). We screened for suppressors of b
d [Su(bd)s] in an effort to isolate mutations that specifically
suppress trans-inactivation. The majority of suppressors recovered
were Su(var)s similar to those previously found to suppress
cis-inactivation of wm4. Surprisingly, nearly a fifth of our Su(bd)
lines carried rearrangements that modified PEV of bw by changing
the chromosomal position of the b d allele, suggesting that this
PEV phenotype depends on nuclear position.
brOwnDominanl (bd). This PEV mutation is unique in
MATERIALS AND METHODS
Fly stocks and culture conditions: Fly stocks were grown on
standard corn meal-molasses medium in shell vials or on instant
food (Carolina Biological Supply) in plastic specimen bottles at
room temperature except as noted. No difference in phenotype was
observed between genotypically identical flies grown on the two
kinds of food.
The b d allele has an insertion of heterochromatin which is
visible as an extra polytene band just proximal to 59E (SLATIS
1955). As it presently exists, b d has the bw gene interrupted by
sequences that lack tested restriction sites (HENIKOFF et al. 1993)
and that are likely to be simple sequence repeats characteristic of
heterochromatin. The gene has no detectable bw+ activity (SLATIS
1955; DREESEN
-
56 1
FIGURE 1.-Eye pigmentation in wild-type and variegating bw geno-
types. Clockwise from upper left:
bup/bw'; st, E(b4)144 b4/+ bw+; st, h+/bw+; st, Szl(b4)20 b4/+
bw+; st,
bWDlb4; st.
et al. 1988). The allele appears to have changed since its
isolation, as HINTON and GOODSMITH (1 950) were able to revert it
to wild type, whereas the current allele appears to be
nonrevertable (K. LOUCHNEY, unpublished data).
The P[bw+]92C strain carries a bw+ transposon inserted 92C
(DREESEN et al. 1991). Its variegating derivative T(2:3)V21e, P[bw+
/" ("V218") was previously described (HEN- IKOFF et al. 1993).
Other mutations not reported here are described by LINDSLEY and ZI"
(1 992).
Mutagenesis and screen: We screened for dominant sup- pressors
of the variegated eye pigment phenotype of bP/+ flies (Figure 1).
The production of pteridine eye pigments by the activity of the bw
locus is more easily monitored in
the absence of ommochrome pigments. The screen and subsequent
analyses were therefore conducted in a scarlet (st) background,
since st+ activity is necessary to produce ommochromes in the eye.
The marker speck (SF) is closely linked to bw (1 -2 cM) and was
sometimes used to distinguish bw+ from bu?' regardless of eye
phenotype (see below).
White-eyed b d sp+; st males were mutagenized with ethyl methane
sul honate (EMS) (GRIGLIATTI 1986) and mass- mated to bw P s t ; st
virgin females (Figure 2). Nearly all of the b d +/+ sp; st progeny
had eyes that were white with a quite uniform scattering of
individual reddish spots within ommatidia (Figure 1, lower right).
Offspring with increased pigmentation ("suppressed flies"; Figure
1, upper right)
-
562 P. B. Talbert, C. D. S. LeCiel and S. Henikoff
white I scarlet F1 b w D + sf [F] b w D + ; st FIGURE 2.-Screen
for suppressors of b d / + ; st.
" x ' s p . s ' + sp Sf + sp Sf + sp Sf
pale yellow variegated orange variegated scarlet
Eye phenotypes are indicated. The nonsuppressed phenotype is
pale yellow variegated (Figure 1, lower right). Individuals with
orange variegated pigmenta-
of about 0.5%. Less than a third of these proved to be fertile
and heritable. The large brackets indicate that the linkage of the
suppressor is unknown.
PmmEl [ -500 (0.5%) I tion (Figure 1, upper right) were
recovered at a rate
t F2 " b w D + , sf
+ sp Sf
pale yellow variegated
[F] -,- b w D + ,sf + sp Sf
orange variegated scarlet scarlet m
99 x T(2;3) V21: bwD Sp; sf Sb P[bw 7 " I bwD sp; ln(3R) P, st
+
orange variegated eyes axillary speck; stubbly bristles; white
eyes with very few pigmented spots
T(2:3) V2 1: bwD sp; sf Sb P[bw +] " T(2;3) V21: bwD sp:sfSb
P[bw+]" FIGURE 3.-Test for suppression of cis-inacti- vation of
V21'. The dominant markers s@+ and Sb
+ - +
su - .L
t si
bwD sp'/n(3R)P, sf
stubbly bristles; white eyes with very eyes
white axillary speck;
few pigmented spots stubbly bristles; pale yellow variegated
eyes
r - - - - . - - - - : orange , variegated eyes? I test class) L
_ I " " " _ J variegated eyes
stubbly bristles; 1 white axillary speck; eyes stubbly
bristles;
orange
were used to establish selection stocks by backcrossing them to
bw+ sf; st flies and selecting the suppressed b d +/+ sp; st
progeny. In addition to the suppressed flies, two flies with
phenotypic enhancement (Figure 1, lower left) were also recovered
from the screen and used to make analogous selection stocks.
In the course of propagating the selection stocks, many of the
suppressors and both enhancers showed linkage to bw", as determined
by the recovery of only a few or no nonsuppressed or nonenhanced b
d +/+ sp; st progeny. These linked suppressors and enhancers were
subsequently balanced over Zn(2LR) SMI, Cy in a st background
unless they were sterile or lethal in one sex. As the Cy-bearing
chromosomc? enhances PEV (SPOFFORD 1976 and our un- published
data), the variegation phenotype of these balanced flies was always
checked by crossing to the bw+ sf chromo- some.
Test for suppression of &-inactivation: Since the b d allele
is null, it cannot be used to test for cis-inactivation. Therefore,
suppressors were tested for their ability to sup- press the
cis-inactivation of the P[bw+] gene at 92C in V21'.
Suppressor-bearing b d / + females were crossed to V21'- bearing
males as described in Figure 3. Some of the sup-
-.- + SP ' sf are closely linked to b4 and V21', respectively,
bwD sp ' / n ( 3 ~ ) p , st and can be used to identify the latter
mutations
independently of their eye pigmentation pheno- axillary speck;
types. pale yellow variegated eyes
axillary speck;
variegated eyes orange
pressors were subsequently tested for suppression of cis-
inactivation using wm4. Suppressor-bearing +/Y; b d / + ; st males
were crossed to wm4; +; + virgin females, and progeny were examined
for the expected class of males suppressed for wm4. Although w n 4
/ K b P / + ; st/+ male progeny could not easily be distinguished
from the wm4/Y; +/+; st/+ male progeny, suppression of wm4 could be
observed in either genotype.
Test for suppression of telomeric variegation: Some suppressors
were tested for their effect on w""; P((w,ry)A] 4-4, a stock in
which a w+ gene inserted at lOOF is subjected to variegation
induced by proximity to the right telomere of the third chromosome
(HAZELRIGG et al. 1984; LEVIS et al. 1985). Suppressed +/Y; b d / +
; st males were crossed to
"s t male progeny (with second chromosomes either b d / + or
+/+) were evaluated for suppression.
Test for suppression of trans-inactivation: P[bw+]92C is
trans-inactivated by its variegating derivative V21" (DREESEN et
al. 1991). Some suppressors of b d / + were tested for their
ability to suppress this trans-inactivation at 92C. Tests were
conducted at 18 a , since this trans-inactivation is more
pronounced at that temperature. Two different crossing
w J 1 1 8 . , P[(w,ry)A]4-4 virgin females. The wJJ18/Y; +
P[(w,ry)A]
-
Modifiers of bp 563
bwD +; st + E] T(2:3) V2 1 bw sp; st Sb P[bw 'I " bwD Sp , st
P[bwf]92C st P[bw+]92C * x % stubbly bristles; orange variegated
eyes scarle7leyes
axilla speck;
1 bwD Sp; st + P[bw']92C I E] T(2;3) V21e, bwD f; st Sb
P[bw']'
+ - bwDSp; St + P[bw']92C + T(2:3) VZC, bwD f; st Sb P[bw+lV
l - - - - - - - - - - 1 stubbly bristles; yellow variegated
eyes
(control)
Su bwD+, St $q X + + w st
orange variegated eyes
/ Su bwD+ st
+ bwDsp' st P[bwf]92C ? ? X
dull scarlet eyes
t
I stubbly bristles; I I red variegated eyes I
I (suppression) I I """"" A
bwD sp . st P[bw'192C bwD sp ' st P[bwf]92C bb
T(23) V27' bwD sp;stSb P[bw+]'
bwD sp: st +
white eyes wlth very axillary speck; stubbly bristles;
few pigmented spots
T(2;3) V21: + bwD sp; st Sb P[bw+]" r - - - - - - - - - - I no
axillary speck; I
1
I stubbly bristles; I yellow variegated eyes I
I
I (no suppression) ' L """"" J FIGURE 4.-Tests for suppression
of trans-inactivation with V21e/
P [ h + ] 9 2 C . The strategies for testing suppressors
unlinked to b4 (A) and for testing suppressors linked to b4 (B)
differ to avoid crossing linked suppressors (which are
rearrangements) onto T(2;3)V21'. The hypomorphic bw+ phenotype
resulting from a single dose of P [ h + ] (DREESEN et al. 1991) is
described as ''dull scarlet." See MATERIALS AND METHODS for
additional details.
strategies were required depending on whether or not the
suppressor was linked to bwD (Figure 4). Unlinked suppres- sors
yielded a class of red-variegated (suppressed) V21'/ P[bw+]92C
progeny, whereas linked suppressors yielded only yellow-variegated
(nonsuppressed) V2lU/P[bw+]92C progeny. To verify that the Su(bd)
was indeed present in these nonsuppressed V21e/P[bw+]92C flies,
individuals were crossed to bw+ sp; st and their progeny
backcrossed to bw+ sp; st to remove the P[bw+]-bearing chromosomes
(not shown). Suppressed b d +/+ sp; st flies reappeared,
indicat-
ing that the Su(bp) was present in the non-suppressed V21'/
P[bw+]92C grandparent.
A single suppressor, Su(bd)44, which was not linked to b p ,
gave inconsistent results in the test for suppression of
cis-inactivation, apparently because the suppression was very weak.
The test for suppression of trans-inactivation with this suppressor
followed the procedure outlined above for linked suppressors,
except that the V21e/P[bw+]92C flies clearly fell into suppressed
and nonsuppressed classes. Individuals in both classes were tested
as described above to verify the presence of the suppressor in
individuals of the former but not the latter class.
Complementation tests: A subset of suppressors, en- hancers,
recombinants and deficiencies were tested for com- mon lethal
mutations in crosses that generally took the form lethal#l/SMl X
lethal#2/SM1. Between 100 and 200 prog- eny were scored for
survival of the lethal#l/lethal#2 progeny except in a few crosses
where it was clear that these progeny were viable after scoring
50-70 progeny. The combinations were classified as lethal if no
lethal#l/lethal#2 progeny were recovered, as semilethal if fewer
than 10% of the expected progeny (as estimated from the sibling
classes) were re- covered, and as viable otherwise.
Pairing tests: Since a large fraction of the suppressors were
rearrangements of chromosome arm 2R that might potentially have
affected pairing between homologs, several 2R chromosome
rearrangements (with bw+) were tested for effects on the eye
phenotype of b d as follows: crosses were performed using
appropriate markers to make b d / + ; st flies carrying the
rearrangements on the bw+ chromosome. In four cases, these were
compared with MI+; st sibs to determine any suppressive effects of
the rearrangements on the eye phenotype. In the remaining two
cases, they were compared with standard b d +/+ sp; st flies.
Cytology: Suppressors and enhancers showing linkage to b d were
examined for rearrangements on the chromosome arm carrying b d
(2R). Suppressor-bearing or enhancer- bearing W/+; st flies were
crossed to bw; st flies, and wandering third-instar larvae were
identified as being b d / bw; st by their colorless (rather than
yellow) Malpighian tubules. These larvae would carry any
rearrangements linked to b d . Salivary glands were dissected in
45% acetic acid and then transferred to a drop of 45% acetic acid
1% orcein on a siliconized coverslip. The coverslip and glands were
transferred to a clean slide and examined on an in- verted
microscope while gently tapping the coverslip with a pencil. When
chromosomes were well spread, they were examined by phase contrast
microscopy.
RESULTS
Trans-inactivation at the bw locus is dependent on pairing
between a heterochromatin-associated bw al- lele and its homolog,
but specific sequences from the bw gene region are not required
adjacent to the het- erochromatic breakpoint (HENIKOFF et al.
1993). To explain this, a model for trans-inactivation was pro-
posed in which a protein essential for bw transcription becomes
inactivated when it binds at the bw locus and is brought into
contact with heterochromatin-binding proteins by somatic pairing of
adjacent homologous sequences. Although it mediates
truns-inactivation, such a protein might be expected to have no
role in cis-inactivation. We reasoned that a mutation in this
hypothetical factor which rendered it insensitive to
-
564 P. B. Talbert, C. D. S. LeCiel and S. Henikoff
contact with heterochromatin, but still permitted bw
transcription, might act as a suppressor specific for
trans-inactivation and be genetically identifiable on this basis.
Although it is not possible to predict whether such a mutation
would be dominant or re- cessive, a dominant mutation would be much
easier to identify both in an initial screen for phenotypic
suppression and in subsequent genetic tests.
Isolation of modifiers: We chose to look for sup- pressors of
the dominant variegating allele bd. This allele is the strongest of
all trans-inactivating bw alleles and is fully viable. This allele
is also null (see MATE- RIALS AND METHODS) and results in a
completely un- pigmented eye in a mutant scarlet background (in
which ommochrome pigments are absent). As a con- sequence,
cis-inactivation of brown in b d is undetect- able and the level of
pteridine production in M/+; st flies directly reflects
trans-inactivation of the bw+ allele. Flies of this genotype
produce pigment in a number of individual ommatidia scattered
throughout the otherwise very pale eye (Figure 1 , lower right).
The consistency of this variegated phenotype among flies allows
sensitive identification of suppressors, al- though the overall
lack of pigment makes enhancers more difficult to detect. We
treated b d ; st males with EMS and crossed them to bw+; st females
(Figure 2). Approximately 100,000 b d / + offspring were screened
to yield 150 dominant suppressor mutations [Su(bd)s] with more eye
pigment than their b d / + ; st siblings (Figure 1, upper right).
With one exception (discussed below), all of the suppressed lines
still had variegated eyes. Thirty-nine of the suppressors either
proved to be too weak to reliably identify in crosses or were lost
prior to analysis. Two dominant enhancer mutations [ E ( b d ) s ]
were also recovered in which flies had fewer pigmented ommatidia
than their b d / + ; st siblings (Figure 1 , lower left).
Suppression of cis-inactivation: The suppressors were tested for
their ability to suppress cis-inactivation of the bw+ gene on the
V21" translocation (Figure 3). This translocation juxtaposes
heterochromatin from the base of 2R to a bw+ gene present within a
P element transposon at 92C on the third chromosome (HENIKOFF et
al. 1993). The bw gene at 92C on the V21' chromosome is strongly
cis-inactivated, yielding an essentially white eye with only a few
pigmented ommatidia in a homozygous b d ; st background. Since this
gene has no paired homolog on a normal (st) third chromosome,
trans-inactivation cannot occur and any phenotypic suppression
observed must be due to the suppression of cis-inactivation.
Of 1 1 1 suppressors tested, 87 suppressed the cis- inactivation
of bw+ on V21", and 24 did not. One of the 87, designated Su(bd)44,
suppressed V21" only weakly and inconsistently and was initially
classified as failing to suppress. It will be discussed further
below.
None of the tested suppressors of cis-inactivation showed tight
linkage to b d , and so are second-site suppressors. Although no
systematic effort was made to map them, in the course of further
testing it became apparent that suppressors were recovered on both
of the large autosomes. A single recessive suppressor, su(bd)62, is
X-linked. Since this suppressor was (nec- essarily) recovered from
a heterozygous female, this female may have been selected because
of a second dominant suppressor in the same fly that subsequently
segregated away and was lost.
To determine whether these suppressors of cis- inactivation were
typical Su(var) mutations or might represent a distinct class of
suppressors of bw varie- gation, 37 of these 87 Su(@)s were tested
for their effect on wm4. All but four suppressed wm4. We there-
fore concluded that about 90% of the suppressors of
cis-inactivation were likely to be typical Su(var)s.
Failure to suppress telomeric variegation: Genes transposed to
Drosophila telomeres are also subject to variegated position
effects (HAZELRIGG et al. 1984). The stock w1'18; P[(w,ry)A]4-4
carries the w+ gene inserted next to the 3R telomere which causes
the gene to variegate. Thirty-four of the 87 Su(bd)s were crossed
to P[(w,ry)A]4-4. All failed to suppress the telomeric variegation,
consistent with tests of other modifiers of wm4 (R. LEVIS,
unpublished data). This suggests that while telomeric variegation
appears phe- notypically similar to PEV caused by pericentric het-
erochromatin, different genes probably are involved in these two
phenomena.
Suppression of trans-inactivation: The 24 Su(bw")s that do not
suppress cis-inactivation in V21" might be expected to include any
suppressors that specifically affect trans- but not
cis-inactivation, as well as any that are allele-specific
suppressors or revertants of b d . To determine if any of the
Su(bd)s are specific suppres- sors of trans-inactivation,
candidates were tested for their ability to increase the amount of
eye pigmenta- tion in V21*/P[bw+]92C flies (see MATERIALS AND
METHODS). P[bw+]92C is the parent allele of V21', and the bw+ genes
on these chromosomes can pair so that V21" acts as a moderately
strong trans-inactivator of P[bw+]92C at 18 O .
As a preliminary positive control, six of the 87 Su(bd)s that
suppress cis-inactivation were tested for suppression of
V2lC/P[bw+]92C flies (Figure 4A); all showed increases in eye
pigmentation consistent with suppression of trans-inactivation.
This was especially clear for Su(bd)44, an unlinked suppressor that
was originally scored as having no effect on cis-inactiva- tion.
Repeated assays revealed a weak effect of Su(bd)44 on
cis-inactivation and a pronounced effect on trans-inactivation. The
stronger effect of Su(bd)44 on expression from the trans copy of bw
than the cis copy suggests that it might be a heterochromatic
com-
-
Modifiers of bp 565
TABLE 1
Suppressors linked to b d
Separable from Suppresses
Suppressor0 b9, V21~lP[bW+]92C? Cytologyb Viable over
Dx2R)aJ?
Su(bwD)5 No T(2;3)55B; 1 OOD Viable Su(bwD)20 Yes No T(1;2)1
lC;52D Su(MY2 Yes No T(1;2)2B;57D Su(bwD)55 No T(1;2)15A;54E
Su(hD)59 No T(2;3)54C;62A Viable Su(bw")73 Yes T(1;2)15E;53F Viable
Su(bwD)87 No No T(Y;2)59E Lethal Su(bwD)98 No ln(2R)59E;60E Viable
Su(bwD)125 No No T(1;2;3)5A;8D;59E;74Cc Su(bwD)126 Yes No
T(2;3)55F;lOOD Su(bwD)131 No In(2LR)26F;59E Viable Su(bwD)133 Yes
No T(2;3)53A;lOOB9 Su(bwD)151 No No T(1;2)17E;57A Viable Su(bwD)l
58 Yes No T(1;2)8E;56E Su(bwD)l 69 No Unrearranged Viable Su(bwD)l
89 No Dj(2R)59E;6OC Viable
a Phenotypes of all suppressors revealed no consistent
differences, except for Su(bwD)169, which is weak, and Su(bwD)189,
which showed
* All suppressors except Su(bwD)189 have a doublet band at 59E
associated with bd'. no variegation. New ;;der uncertain.
. .
ponent that is more directly involved in mediating
trans-inactivation than the other suppressors of cis-
inactivation.
The 24 Su(bwD)s that did not suppress cis-inactiva- tion in the
test with V21a all showed linkage to b d in crosses. Six were
clearly separable from the bw locus, but others were apparently
inseparable, since no un- suppressed b d / + flies were recovered
during stock maintenance (Table 1). Five of the six separable
linked suppressors and three apparently inseparable suppres- sors
were tested for their effects on the phenotype of V21g/P[bw+]92C
flies (Figure 4B). None of these eight S u ( b d ) s suppressed
either cis- or trans-inactivation of the bw+ insert at 92C (Table
1). These results sug- gested that many or all of the Su(bd)s
linked to b d are specific suppressors of the b d allele, since
they do not affect cis- or trans-inactivation when present with
other variegating alleles.
Cytology of suppressors linked to bz8: The exist- ence of at
least 5 Su(bd)s that were linked to but separable from b d and that
were apparently specific suppressors of b d raised the possibility
of an unusual interaction between bwD and a neighboring locus or
cluster of loci. Most of the 24 linked Su(6d)s behaved genetically
like translocations: for example, 11 Su(bwD)s were either
male-lethal or male-sterile, indi- cating that they were probably
translocations between the X and second chromosomes (LINDSLEY
1982). Cytological examination was undertaken to determine whether
breakpoints common to different re- arrangements would identify a
unique suppressor lo- cus. The cytology for the surviving 16 of the
original
24 linked Su(M)s is presented in Table 1. Of the 16 suppressors,
five have breakpoints at 59E,
the site of the bw gene: one of these is a deficiency, two are
inversions, one is a translocation to the Y chromosome, and one is
a complex rearrangement involving three chromosome arms. Su(bwD)I89
deletes the b d doublet at 59E, which explains why this chro-
mosome alone of all the Su(bd)s does not cause de- tectable
trans-inactivation. Su(M)87 segregates ster- ile D m Dp(2R)59E;60F,
bp +/bw+ sp/bw+ sp; st males as well as fertile T(y;2)59E, b d
+/bw+ sp; st males. The former are distinguished by their more
deeply pigmented red-variegated eyes, presumably a result of having
two bw+ genes compared with only one in their less pigmented
fertile brothers. Su(bd)87 and the other three lesions at 59E might
be expected to disrupt pairing at bw, which could account for the
observed phenotypic suppression.
The remaining 1 1 linked suppressors do not show alterations at
59E: one is a weak suppressor that is cytologically
indistinguishable from its b d parent chromosome, six are
translocations to the X chromo- some, and four are translocations
to the third chro- mosome. The distribution of breakpoints in the
latter 10 rearrangements appears to be very nonrandom (Figure 5).
The breakpoints on the second chromo- some are all between 52D and
57D, although no two are in the same place. The six breakpoints on
the X chromosome are scattered from 2B to 17E. In con- trast, the
four breakpoints on the third chromosome are at the distal tips.
Given the clustered but noncoin- cident distribution of breakpoints
on 2R, it seems
-
566 P. B. Talbert, C. D. S. LeCiel and S. Henikoff
Suppressors
2L I
Enhancers 2L 1
II I II
4 + FIGURE 5.-Distribution of translocation breakpoints of
suppres-
sors and enhancers. T(J;2) and T(2;3) suppressors as well as
T(2;het) enhancers all have breakpoints on chromosome arm 2R
clustered between 52D and 57D, proximal to bup at 59E. The
breakpoints on the third chromosome of T(2;3) suppressors are
confined to the distal tips, while the breakpoints on the X
chromosome of T(1;2) suppressors are widely distributed. The
enhancers have breakpoints in heterochromatin: E(b4)40 =
T(2;4)54F;JOIhet and E(bup)144 = T(2;3)54B;8Ohet.
unlikely that any breakpoint represents the position of a gene
that acts as an allele-specific suppressor of M/+. It appears
instead that translocations in a por- tion of chromosome arm 2R
proximal to bp can affect its ability to trans-inactivate its
homolog, and that all such translocations to the X chromosome
affect trans-inactivation while only those at the tips of a large
autosome are able to do so.
One possible way in which a rearrangement might affect
trans-inactivation is by affecting the probability of synapsis of
homologs. If a rearrangement in one member of a chromosome pair
decreases the proba- bility of somatic pairing, we would expect
that such rearrangements would suppress trans-inactivation. We
tested six rearrangements of 2R isolated in other studies for their
possible effects on trans-inactivation by bwD and found none (Table
2). One notable differ- ence between the rearrangements in these
tests and the Su(bp)s, other than breakpoint position, is that the
rearrangements listed in Table 2 are all on the bw+ chromosome,
while the Su(bp)s are all re- arrangements on the bwD chromosome.
While this difference in linkage should not affect pairing, it
might affect other processes such as nuclear position- ing or
heterochromatin formation at bp.
Enhancers: The distribution of rearrangement breakpoints at the
distal tips of chromosome arms 3L and 3R in Su(bwD)5, Su(bwD)59,
Su(bp)I26 and Su(bp)133 suggests that the placement of the bwD
heterochromatic block distally, i e . , farther from the
pericentric heterochromatin, may reduce the effi- ciency of
trans-inactivation. This leads to the expec-
TABLE 2
2R rearrangements tested for suppression of
trans-inactivation
Suppresses Rearrangement Cytology bg/+; Sf?
In(2R)AA21 + Dfl2R)AAZJ In(2R)56E;58E + No DfT2R)56F;57D
w2R)vgu In(2R)49C;5OC No Tp(2;y)bw+ Tp(2;y)58F- Noa
T(2;3)Antpd T(2;3)4 1 F;84B No T(2;3)C287 T(2;3)56D;89F No
T(2;3)Ta’ T(2;3)5lE;84B No
a Data from (HENIKOFT and DREESEN 1989).
59A;60EF
A.
FIGURE 6.-Association of bup with the chromocenter in a
7‘(2;3)54B;80het. E(b4)144 b4/+bw; st salivary gland nucleus. The
entire synapsed right arms of the second chromosomes are visible.
The bw chromosome arm is continuous while the E(bup)J44 bup arm is
interrupted by the translocation at 54B and is associated with the
heterochromatic chromocenter at 59E, the location of b4.
tation that, conversely, enhanced trans-inactivation will be
observed when the bp heterochromatic block is moved proximally, i e
. , closer to the pericentric heterochromatin. This is precisely
what was seen for the two E ( b p ) s isolated in the screen for
mutations.
Cytological examination showed that both E(bp)s were associated
with translocations of 2R to het- erochromatin: E ( b p ) 4 0 is
T(2;4)54F;IOlhet and E(bwD)144 is T(2;3)54B;80het. It is
interesting that both E(bup) breakpoints on 2R are in the same
region as the 2R breakpoints of the Su(bp)s. Enhancement cannot be
explained by extreme heterochromatic spreading, since the distance
from the breakpoint to bwD is several-fold greater than has been
previously observed (SPOFTORD 1976), and since the banding pattern
of the intervening euchromatin is unaffected (Figure 6). An
alternative explanation for enhance- ment is suggested by a
striking feature of the polytene chromosome cytology of these
enhancers: the bwD heterochromatic block at 59E is frequently
associated with the chromocenter in both stocks (Figure 6), al-
though it is almost never associated with the chrom- ocenter in
bp/+ (Table 3). Since the two enhancers have different
translocation breakpoints, the associa- tion of 59E with the
chromocenter is evidently
-
Modifiers of b P 567
TABLE 3
Association of a4 with heterochromatin of salivary
chromosomesa
Distance of b9 from
heterochromatin Nuclei with Nuclei with (no. of lettered b9 at
b9 not at
Genotype subdivisions) chromocenter chromocenter
bwD/+ 113 1 (2%) 50 T(2;4) E(bwD)40/bw 29 29 (78%) 8 E ( b ~ ~ )
4 ~ ~ / b w 113 0 (0%) 42 T(2;3) E(bwD)144/bw 33 37 (86%) 6
E(b4)144REc/bw 113 4 (8%) 45 T(2;3) Su(bwD)5/bw 143 0 (0%) 74
T(1;2) Scl(bwD)73/bw 67 10 (21%) 38 T(1;2) Su(b4)151/bw 35 23b
(66%) 12
,I Based on combined data for scorable nuclei from two squash
preparations per genotype.
For all nine nuclei in which X-heterochromatin had separated
from the bulk of the chromocenter, b w D was associated with X, not
autosomal heterochromatin.
brought about by the proximity of b d to the hetero- chromatic
breakpoints, and not to disruption of a gene on 2R euchromatin.
Evidence that the association of b d with the chrom- ocenter of
salivary nuclei correlates with enhancement of PEV is provided by
recombinant derivatives of the translocations. The recombinant
E(bd)144REC was re- covered from our E(bd)144 b d +/+ + sp
selection stock by virtue of its nonenhanced phenotype, which was
indistinguishable from that of b d / + . A sim- ilar recombinant, E
( b d ) 4 P c , was obtained from E(bwD)40 b d +/+ + sp. E ( b d )
4 P c is semilethal as a homozygote and shares its semilethality
with E ( b d ) 4 0 , verifying its derivation from this enhancer.
Cytologi- cal examination revealed that in both recombinants the b
d heterochromatic block had recombined away from the translocation
onto the normal second chro- mosome, and that 59E is no longer
associated with the chromocenter (Table 3). Furthermore, both E ( b
d ) 4 0 and E(bd)144 are semilethal over the small deficiency,
Dfl2R)bw'. Semilethality in each case can be explained as enhanced
cis-inactivation of a vital gene uncovered by the deficiency. This
enhancement of cis-inactivation does not occur in the viable recom-
binant E(bd)144REC, nor in E ( b d ) 4 p c , which is via- ble over
Df12R)bw5, even though the bw region should be identical in these E
( b d ) s and the recombinants derived from them. Therefore
enhancement of both cis-inactivation of a vital locus and
trans-inactivation of bw' is relieved by separating b d from a
transloca- tion breakpoint that causes it to associate with the
chromocenter in salivary nuclei.
An example of a suppressor with a distal breakpoint, Su(bd)5 ,
showed no association between b d and the chromocenter of salivary
gland chromosomes (Table 3). However, three of the bd-specific
suppressors that translocate b d proximally did exhibit an
association.
Su(bd)87 places band 59E1-2 adjacent to the hetero- chromatic Y
chromosome. It is possible that the le- thality of Su(bd)87 over
Df12R)bw5 (Table 1) is due to enhancement of cis-inactivation in
this stock: the jux- taposition to pericentric heterochromatin may
cause spreading to occur into a neighboring essential gene.
However, the pairing of b d with its homolog is visibly disrupted
in most Su(bd)87 salivary gland nuclei (P. B. TALBERT, unpublished
data). A similar pairing disruption in the pigment cell nuclei
would lead to suppression of trans-inactivation. Su(bp)73 and
Su(bd)151 translocate bwD very proximally onto the X chromosome
(15E and 17E, respectively), leading to association of b d with the
base of the X chromo- some in salivary gland nuclei (Table 3). In
many of these nuclei, the base of the X (and b d ) is separated
from the rest of the chromocenter. If this feature parallels a
difference in the position of the X chro- mosome in pigment cells
of the eye, it might account for the lack of enhancement in these
rearrangements.
DISCUSSION
A screen for modifiers of b d position-effect varie- gation led
to the isolation of an unexpected class of mutations:
translocations that alter the linkage of the b d heterochromatic
element without affecting the element itself. These translocations
all have one breakpoint within a region that is proximal to but
separable from b d , and have the other breakpoint at a position
that depends upon whether the chromo- some involved is the X or an
autosome. Recovery of this unexpected class is accounted for by
differences between our screen, designed to obtain specific dom-
inant suppressors of trans-inactivation, and earlier studies.
Search for specific suppressors of trans-inactiva- tion: We
sought mutations in a hypothetical hetero- chromatin-sensitive
transcription factor that specifi- cally suppress
trans-inactivation. The desired muta- tions would retain bw
transcriptional activity while exerting dominant insensitivity to
the presence of heterochromatin. None of the suppressors strictly
ful- filled the criteria we imposed to identify such a mu- tation:
that it have no effect on cis-inactivation but be able to
dominantly suppress trans-inactivation at an ectopic site (92C) as
well as at the bw locus. It may be that this combination of
characteristics is very improb- able for mutations in the
hypothetical transcription factor. For example, suppressors
specific for trans- inactivation may only be recoverable as
recessives. Alternatively, it may be that the hypothetical tran-
scription factor does not exist and trans-inactivation is a direct
interaction between the primary sequence or secondary structure of
DNA in or near the bw gene and heterochromatic proteins brought
near by pairing between homologs. Sequence analysis of the bw
gene
-
568 P. B. Talbert, C . D. S . LeCiel and S. Henikoff
has not revealed any repetitive or unusual sequences that would
make obvious targets for heterochromatin- forming proteins
(MARTIN-MORRIS et al. 1993). The resolution of this question may be
helped by further characterization of the bw sequences necessary to
me- diate trans-inactivation.
Unlinked suppressors: The Su(bd)s can be grouped into two large
classes based on their linkage to the bw locus and their ability to
suppress cis-inacti- vation of V 2 P . The majority of suppressors
are clas- sical Su(uar)s, typically unlinked to b w , which
generally suppress &inactivating PEV mutations. Thus, despite
its especially strong dominant effect caused by an interstitial
insertion of heterochromatin, it is clear that b d is in many
respects a typical PEV mutation. Pre- vious screens for Su(uur)s
and E[var)s utilizing wm4 have failed to identify any Su(uar)s on
the X chromosome. While our results are generally consistent, since
we have not identified any dominant suppressors on the X
chromosome, we did recover an X-linked recessive suppressor of
variegation. It has previously been noted (GRIGLIATTI 1991) that a
minority of Su(uar)s show some allele-specificity and this also may
be true of about 10% of the Su(bd)s. None of the 34 tested Su(bd)s
had any effect on telomeric variegation, in- dicating that
different components are responsible for variegation caused by
telomeric and centromeric re- gions.
Linked suppressors: The second large class of Su(bd)s consists
of 24 suppressors linked to 6 2 8 itself. All of these failed to
suppress the cis-inactivation of V21', and eight that were tested
also failed to suppress trans-inactivation by this rearrangement,
indicating that they are allele-specific suppressors of b d . All
but one of the 16 examined cytologically proved to have
rearrangements on 2R. Since EMS is thought to in- duce mutations
primarily by base substitution (COTE et al. 1986), the almost
perfect correlation between rearrangements linked to b d and
allele-specific suppression is compelling evidence that it is the
struc- tural alterations of the chromosome arm carrying b d rather
than particular loci disrupted by the re- arrangements that lead to
allele-specific suppression.
These bd-linked rearrangements can be further subdivided into
three groups: rearrangements at 59E, translocations to the X
chromosome, and transloca- tions to the distal tips of the third
chromosome. The first type of rearrangement was recovered by HINTON
and GOODSMITH (1 950) in their reversion study of b d using
X-irradiation. Rearrangements at 59E would be expected to disrupt
pairing locally, preventing trans- inactivation in some cells. They
might also dissect the b d heterochromatic block into two pieces,
altering its ability to induce variegation.
The other two classes of rearrangements were not recovered by
HINTON and GOODSMITH. The de-
creased sensitivity to phenotypic changes afforded by using a
st+ background may account for their failure to recover them. In
addition, the b d allele appears to have changed structurally since
1950 (see MATE- RIALS AND METHODS), and its phenotype may have been
less sensitive to this type of rearrangement at that time.
Breakpoint distribution: Translocation break- points associated
with b d / + suppressors and en- hancers are strikingly nonrandom
(Figure 5). All 12 breaks proximal to b d are clustered within a
region consisting of about 20% of chromosome arm 2R. If the entire
arm proximal to b d is considered to be a potential target, then
the chance probability of such clustering is about (0.2)'' lo-'.
The autosomal breakpoints associated with suppressors are also non-
randomly distributed along 3L and 3R, the two au- tosomal arms
represented. All four breakpoints lie within the distalmost 5% or
so of each arm, a distri- bution expected by chance to occur at P =
(0.05)4 =
The rarity of these autosomal breakpoints rela- tive to the six
scattered breakpoints on the X chro- mosome is consistent with the
assumption that the large majority of possible autosomal
translocations failed to lead to detectable phenotypic change and
so were not selected in the screen. This assumption seems
reasonable, since we are unaware of evidence for restricted
occurrence of chromosomal rear- rangements following mutagenesis
(ASHBURNER
The clustering of Su(bd) and E ( b d ) translocation breakpoints
between 52D and 57D might suggest that this region is important for
initiating synapsis of the 2R chromosome arms, since
trans-inactivation is a somatic pairing-dependent phenomenon
(HENIKOFF et al. 1993). However, trans-inactivation is not the only
bd-associated phenotype affected by linkage altera- tions: each of
the E ( b d ) s is also enhanced for cis- inactivation of a nearby
essential gene, which is not expected to be sensitive to somatic
pairing disruptions. In addition, it is difficult to explain why
translocations that move the trans copy of bw, such as T(Y;Z)bw+Y
and T(2;3)C287 (Table 2) have no effect. Other explana- tions for
this puzzling clustering of breakpoints must be considered, such as
a special position within the nucleus. In this regard, we note that
the only signifi- cant point of association between the nuclear
envelope and salivary chromosome arm 2R proximal to b d is within
the 52D-57D region (HOCHSTRASSER et al. 1986; MATHOG and SEDAT
1989).
The distal translocation breakpoints on the third chromosomes of
the T(2;3) suppressors together with the heterochromatic
breakpoints in the enhancers suggest that distance from the
chromocenter is an important determinant of heterochromatin
formation on the autosomes. The T(2;3) suppressors increase the
1990).
-
Modifiers of b d 569
length of the chromosome arm bearing b p , while the enhancers
greatly decrease it. The proximity of 67.8 to the centromere in the
enhancers leads to its asso- ciation with the chromocenter of
salivary nuclei (Fig- ure 6). Although we do not know the nuclear
location of b d in pigment cells, the correlation between its
chromocentral location in salivary nuclei and the en- hancement of
both cis-inactivation and truns-inactiva- tion in both E ( b p ) s
is consistent with a similar local- ization in pigment cells. This
localization may facilitate heterochromatin formation in the b d
heterochro- matic block, leading to the observed enhancement. The
T(2;3) suppressors may make heterochromatin formation at the b d
heterochromatic block more difficult by moving it farther than
normal from the chromocentral compartment. It should be noted,
however, that the linear distance along the chromo- some arm need
not accurately reflect the three-dimen- sional spatial relationship
between b d and a chro- mocentral compartment in pigment cells. For
exam- ple, it is possible that the existence of Su(bwD) break-
points throughout the X euchromatin is accounted for by a special
compartmentalization or loose association of the X chromosome,
which places it effectively “far- ther’’ from the chromocentral
compartment.
The abundance of X-linked suppressors perhaps reflects a large
target size for translocation breaks in X euchromatin relative to
the autosomal tips. Of the 24 linked suppressors recovered, 11
showed male lethality or male sterility indicative of a T(2;2)
trans- location, and this conclusion was confirmed for all seven of
those that survived long enough for cytolog- ical examination.
These results are reminiscent of the findings of KHVOSTOVA (1939)
for the cubitus inter- ruptus (ci) locus, which is located in the
vicinity of the heterochromatin-euchromatin junction of chromo-
some 4. Normally, ci+ is dominant over the ci’ allele. However,
translocations of ci+ show reduced domi- nance when the locus is
moved to distal but not to proximal regions of the autosomal arms.
In addition, translocations of ci+ to all euchromatic portions of
the X chromosome similarly showed reduced dominance. So like the
reduced dominance of b7.8 (over bw+) observed for suppressors
obtained in our screen, re- duced dominance of ci+ (over ci’) was
seen for trans- locations both to distal autosomal sites and to
any- where on the euchromatic X. We suggest that both phenomena
have a similar causal basis. Furthermore, since breaks proximal to
ci+ are likely to be in heter- ochromatin, the repositioning of ci+
with adjacent heterochromatin to euchromatic regions resembles the
repositioning of the b d heterochromatic element described
here.
The results of KHVOSTOVA (1939) and those of PANSHIN (1938) also
indicate that the distal hetero- chromatin of the sex chromosomes
differs from other
heterochromatin. In particular, KHVOSTOVA observed that X
heterochromatin distal to the bobbed locus, the site of the
nucleolus organizer, behaved similarly to X or distal autosomal
euchromatin in its ability to reduce the dominance of ci+, but X
heterochromatin proximal to the bobbed locus behaved like autosomal
hetero- chromatin. This suggests that the nucleolus might form a
boundary between two distinct kinds of het- erochromatin, the more
distal of which resembles euchromatin in some properties. This may
be relevant to the failure of Su(b7.8)s broken in proximal X eu-
chromatin to enhance variegation: the observed asso- ciation of
b7.8 with the (sometimes displaced) base of the X chromosome in
these suppressors may involve only the distal “euchromatic-like”
heterochromatin.
Relationship to models for position-effect varie- gation: Early
studies of position effects on genes lo- cated in proximal regions
led to the notion that blocks of heterochromatin are important for
the functioning of heterochromatic genes (KHVOSTOVA 1939; LEWIS
1950; BAKER 1953; HESSLER 1958). Variegation of heterochromatic
genes has been hypothesized to re- flect their association with a
chromocentral compart- ment in some cells but not others (WAKIMOTO
and HEARN 1990). These position effects depend on the distance of
heterochromatic genes from large heter- ochromatic blocks, not the
centromere itself (PANSHIN 1938; EBERL et al. 1993). Furthermore,
heterochro- matic regions may need to aggregate to function
(WAKIMOTO and HEARN 1990; EBERL et al. 1993).
Our results provide complementary support for these ideas, in
that the ability of a block of hetero- chromatin to inactivate
eushromatic genes also ap- pears to depend on its proximity to
other heterochro- matin. This proximity might be necessary for
access to heterochromatin-binding proteins localized in a
chromocenter (WAKIMOTO and HEARN 1990). Al- though no association
of b d with other heterochro- matic blocks is observed in b d / +
polytene salivary glands, we suggest that this association occurs
in the thinner and presumably more flexible diploid chro- mosomes
of the pigment cells. The probability of this association, and
therefore of cis- and truns-inactiva- tion, may be strongly
dependent on the distance of bwD from the heterochromatic
pericentric region as is observed in salivary nuclei of the
E(bwD)s. The prob- ability of association might be decreased in the
X- linked Su(bd)s because of the presence in X-hetero- chromatin of
the nucleolus, which is suggested to impede access of b d to the
chromocenter (A. HILLI- KER, personal communication). In the same
manner, the nucleolus might interfere with the positioning of ci+
in X 4 translocations, thus accounting for the ability of all
breakpoints distal to the nucleolus to reduce the dominance of ci+
over ci’ (KHVOSTOVA 1939).
Chromosome looping into a heterochromatic com-
-
570 P. B. Talbert, C. D. S. LeCiel and S. Henikoff
partment on a smaller scale could underlie the phe- nomenon of
heterochromatic “spreading.” Faculta- tively susceptible
euchromatic sequences might be recruited into the heterochromatic
compartment by their close proximity. Such a model of spreading
could explain the unusual cases of “discontinuous compac- tion”
observed by BELYAEVA and ZHIMULEV (1991), in which polytene
chromosome bands adopted a het- erochromatin-like appearance even
though separated from heterochromatin by normally appearing bands.
Perhaps in these cases, the heterochromatin-like bands are present
in the heterochromatic compartment but are displaced when squashed
preparations are made for cytological analysis.
This model might also account for an apparent discrepancy
between our results (as well as those of R. LEVIS, unpublished
data), which show that Su(var)s fail to suppress variegation of a
w+ gene inserted into the 3R telomere, and those of KARPEN and
SPRADLING (1992), which show that Y chromosomes indeed sup- press
variegation of a rosy gene inserted into the telomere of a
mini-chromosome, Dp(l;f l l l87. Since the telomere of Dp(l;f l l
l87 is only about 250 kb from the pericentric heterochromatin
(KARPEN and SPRA- DLING 1992), we suggest that the observed rosy
varie- gation results from frequent looping into the chrom- ocenter
where the gene becomes sensitive to PEV modifiers. In contrast, the
3R telomere is expected to be located at the opposite side of the
nucleus (HIR- AOKA et al. 1993), where it would be unable to make
contact with the chromocenter, and this difference would account
for the insensitivity of the w+ gene to PEV modifiers. As is the
case for the linkage altera- tions of b d described here,
differences in variegating phenotypes would thus depend upon the
relative lo- cations of the chromocenter and smaller genetic ele-
ments (telomeres) which are rich in repetitive se- quences (KARPEN
and SPRADLING 1992; R. LEVIS, unpublished data).
We thank JEFF JACKSON and ADRIAN QUINTANILLA for technical
assistance and stock maintenance, and our many colleagues who
provided critical comments on the manuscript. S. H. thanks GARY
KARPEN and ALLAN SPRADLINC for discussions of the effects of
Su(var)s on telomeric variegation and ART HILLIKER for pointing out
that the nucleolus might impede associations between hetero-
chromatic elements located on either side. This work was supported
by grants to S. H. from the National Science Foundation
(DCB8717937) and the National Institutes of Health (GM29009).
LITERATURE CITED
ASHBURNER, M., 1990 Drosophila, A Laboratory Handbook. Cold
Spring Harbor Press, Cold Spring Harbor, N.Y.
BAKER, W. K., 1953 V-type position effects of a gene in
Drosophila uirilis normally located in heterochromatin. Genetics 3
8 328- 344.
BAKER, W. K., 1968 Position-effect variegation. Adv. Genet. 1 4
133-169.
BELYAEVA, E. S., and I. F. ZHIMULEV, 1991 Cytogenetic and
molecular aspects of position effect variegation in Drosophila
111. Continuous and discontinuous compaction of chromosomal
material as a result of position effect variegation. Chromosoma
COTE, B., W. BENDER, D. CURTIS and A. CHOVNICK, 1986 Molecular
mapping of the rosy locus in Drosophila mel- anogaster. Genetics
112: 769-783.
DREESEN, T. D., S. HENIKOFF and K. LOUGHNEY, 199 1 A pairing-
sensitive element that mediates trans-inactivation is associated
with the Drosophila brown gene. Genes Dev. 5: 331-340.
DREESEN, T. D., D. H. JOHNSON and S. HENIKOFF, 1988 The brown
protein of Drosophila melanogaster is similar to the white protein
and to components of active transport complexes. Mol. Cell. Biol.
8: 5206-5215.
DUBININ, N. P., 1936 A new type of position effect.
Biologicheskij Zhurnal5: 851-874.
EBERL, D. F., B. J. DUYF and A. J. HILLIKER, 1993 The role of
heterochromatin in the expression of a heterochromatic gene, the
rolled locus of Drosophila melanogaster. Genetics 134 277- 292.
EISSENBERG, J. C., 1989 Position effect variegation in
Drosophila: towards a genetics of chromatin assembly. Bioessays 11:
14- 17.
EPHRUSSI, B., and E. SUTTON, 1944 A reconsideration of the
mechanism of position effect. Proc. Natl. Acad. Sci. USA 30:
183-197.
GATTI, M., and S. PIMPINELLI, 1992 Functional elements in Dro-
sophila melanogaster heterochromatin. Ann. Rev. Genet. 26
239-275.
GRIFFEN, A. B., and W. S. STONE, 1940 The Wm5 and its deriva-
tives. University of Texas Publication 4030: 190-200.
GRIGLIATTI, T., 1986 Mutagenesis, pp. 39-58. In: Drosophila, a
Practical Approach, edited by D. B. ROBERTS. IRL Press, Ox-
ford.
GRICLIATTI, T., 199 1 Position-effect variegation-An assay for
nonhistone chromosomal proteins and chromatin assembly and
modifying factors, pp. 587-627. In: Functional Organization of the
Nucleus: A Laboratory Guide, edited by B. A. HAMKALO and S. C. R.
ELGIN. Academic Press, San Diego.
HARTMANN-GOLDSTEIN, I . J., 1966 Relationship of heterochro-
matin to puffs in a salivary gland chromosome of Drosophila.
Naturwissenschaften 53: 9 1.
HAYASHI, S., A. RUDDELL, D. SINCLAIR and T . GRIGLIATTI, 1990
Chromosomal structure is altered by mutations that suppress or
enhance position effect variegation. Chromosoma
HAZELRICC, T. , R. L E V I S ~ ~ ~ G . M. RUBIN, 1984
Transformation of white locus DNA in Drosophila: dosage
compensation, zeste interaction, and position effects. Cell 36:
469-48 1.
HEARN, M. G., A. HEDRICK, T . A. GRICLIATTI and B. T. WAKI-
MOTO, 1991 The effect of modifiers of position-effect varie- gation
on the variegation of heterochromatic genes of Drosoph- ila
melanogaster. Genetics 128: 785-797.
HENIKOFF, S., 1981 Position-effect variegation and chromosome
structure of a heat shock puff in Drosophila. Chromosoma (Berl.)
83: 381-393.
HENIKOFF, S., 1990 Position-effect variegation after 60 years.
Trends Genet. 6: 422-426.
HENIKOFF, S., and T. D. DREESEN, 1989 Trans-inactivation of the
Drosophila brown gene: evidence for transcriptional repression and
somatic pairing dependence. Proceedings of the National Academy of
Sciences of the United States of America 8 6 6704- 6708.
HENIKOFF, S., K. LOUGHNEY and T. D. DREESEN, 1993 The enigma of
dominant position-effect variegation in Drosophila, pp. 183-196.
In: The Chromosome, edited by J. S. HESLOP- HARRISON. BIOS,
Oxford.
HESSLER, A., 1958 V-type position effects at the light locus
in
100: 453-466.
9 9 39 1-400.
-
Modifiers of b P 57 1
Drosophila melanogaster. Genetics 43: 395-403. HILLIKER, A. J.,
and D. G. HOLM, 1975 Genetic analysis of the
proximal region of chromosome 2 of Drosophila melanogaster. I.
Detachment products of compound autosomes. Genetics 81:
HINTON, T., and W. GOODSMITH, 1950 An analysis of phenotypic
reversions at the brown locus in Drosophila. J. Exp. Zool. 114
HIRAOKA, Y., A. S. DERNBURG, S. J. PARMELEE, M. C. RYKOWSKI, D.
A. AGARD and J. W. SEDAT, 1993 The onset of homolo- gous chromosome
pairing during Drosophila melanogaster em- bryogenesis. J. Cell
Biol. 120 591-600.
HOCHSTRASSER, M., D. MATHOG, Y. GRUENBAUM, H. SAUMWEBER and J.
W. SEDAT, 1986 Spatial organization of chromosomes in the salivary
gland nuclei of Drosophila melanogaster. J. Cell Biol. 102:
112-123.
KARPEN, G., and A. C. SPRADLING, 1990 Reduced DNA polyten-
ization of a minichromosome region undergoing position-effect
variegation in Drosophila. Cell 63: 97-107.
KARPEN, G. H., and A. C. SPRADLING, 1992 Analysis of subtelom-
eric heterochromatin in the Drosophila minichromosome D p l l 8 7
by single P element insertional mutagenesis. Genetics
KAUFMANN, B. P., 1942 Reversion from roughest to wild-type in
Drosophila melanogaster. Genetics 27: 537-549.
KHVOSTOVA, V. V., 1939 The role played by the inert chromo- some
regions in the position effect of the cubitus interruptus gene in
Drosophila melanogaster. Izv. Akad. Nauk SSSR Otd. Ser. Biol. 4:
541-574.
LAIRD, C. D., W. Y. CHOOI, E. H. COHEN, E. DICKSON, N. HUTCH-
INSON, et al . , 1973 Organization and transcription of DNA in
chromosomes and mitochondria of Drosophila. Cold Spring Harbor
Symp. Quant. Biol. 38: 3 1 1-327.
LEVIS, R., T. HAZELRIGG and G. M. RUBIN, 1985 Separable cis-
acting control elements for expression of the white gene of
Drosophila. EMBO J. 4: 3489-3499.
LEWIS, E. B., 1950 The phenomenon of position effect. Adv.
Genet. 3: 73-1 15.
LINDSLEY, D. L., 1982 The genetics of male fertility. Drosophila
Information Service 58: 2.
LINDSLEY, D. L., and G. G . ZIMM, 1992 The genome of
Drosophila
LOCKE, J., M. A. KOTARSKI and K. D. TARTOF, 1988 Dosage-
705-721.
103-1 14.
132: 737-753.
melanogaster. Academic Press, San Diego.
dependent modifiers of position effect variegation in Drosophila
and a mass action model that explains their effect. Genetics
MARTIN-MORRIS, L. E., K. LOUGHNEY, E. 0. KERSHISNIK, G. POOR-
TINGA and S. HENIKOFF, 1993 Characterization of sequences
responsible for trans-inactivation of the Drosophila brown gene.
Cold Spring Harbor Symp. Quant. Biol. 5 8 (in press).
MATHOG, D., and J. W. SEDAT, 1989 The three-dimensional or-
ganization of polytene nuclei in male Drosophila melanogaster with
compoundXYor ring X chromosomes. Genetics 121: 293- 311.
PANSHIN, I. B., 1938 The cytogenetic nature of the position
effect of the genes white (mottled) and cubitus interruptus.
Biologi- cheskij Zhurnal7: 837-868.
POKHOLKOVA, G. V., I. V. MAKUNIN, E. S. BELYAEVA and I. F.
ZHIMULEV, 1993 Observations on the induction of position effect
variegation of euchromatic genes in Drosophila melano- gaster.
Genetics 133: 231-242.
RABINOWITZ, M., 1941 Studies on the cytology and early embryol-
ogy of the egg of Drosophila melanogaster. J. Morphol. 6 9 1-
49.
REUTER, G., and P. SPIERER, 1992 Position effect variegation and
chromatin proteins. Bioessays 14: 605-612.
RUDKIN, G. T., 1969 Non-replicating DNA in Drosophila. Ge-
netics 61 (Suppl.): 227-238.
SLATIS, H. M., 1955 A reconsideration of the brown-dominant
position effect. Genetics 4 0 246-251.
SPOFFORD, J. B., 1976 Position-effect variegation in Drosophila,
pp. 955-1019. In: Genetics and Biology of Drosophila, edited by M.
ASHBURNER and E. NOVITSKI. Academic Press, London.
SPRADLING, A. C., and G. H. KARPEN, 1990 Sixty years of mystery.
Genetics 126 779-784.
TARTOF, K. D., C. HOBBS and M. JONES, 1984 A structural basis
for variegating position effects. Cell 37: 869-878.
UMBETOVA, G. H., E. S. BELYAEVA, E. M. BARICHEVA and I. F.
ZHIMULEV, 199 1 Cytogenetic and molecular aspects of posi- tion
effect variegation in Drosophila melanogaster IV. Under-
replication of chromosomal material as a result of gene inacti-
vation. Chromosoma 101: 55-61.
WAKIMOTO, B. and M. HEARN, 1990 The effects of chromosome
rearrangements on the expression of heterochromatic genes in
Chromosome 2L of Drosophila melanogaster. Genetics 125:
141-151.
120: 181-198.
Communicating editor: A. CHOVNICK