Download P-Element Transformation with period Locus DNA Restores

Cell, Vol. 39, 369-376,
December
1984 (Part I), Copyright
0 1984 by MIT
0092.8674/84/120369-08
$02,00/O
P-Element Transformation
with period Locus
DNA Restores Rhythmicity to Mutant,
Arrhythmic Drosophila melanogaster
\
William A. Zehring,* David A. Wheeler,*
Pranhitha Reddy,T-Ronald
J. Konopka;*
Charalambos
P. Kyriacou,”
Michael Rosbash,* and Jeffrey C. Hall*
* Department of Biology
+ Department of Biochemistry
Brandeis University
Waltham, Massachusetts 02254
* Department of Biology
Clarkson College
Potsdam, New York 13676
9 Department of Genetics
University of Leicester
Leicester LEI 7RH, England
Summary
Mutations
at the period (Per) locus of Drosophila
melanogaster
disrupt
several
biological
rhythms.
Molecular cloning of DNA sequences
encompassing
the per+ locus has allowed germ-line
transformation
experiments
to be carried out. Certain subsegments
of the per region, transduced
into the genome
of
arrhythmic per“ flies, restore rhythmic@
in circadian
locomotor
behavior and the male’s courtship
song.
Introduction
Biological rhythms are ubiquitous and have been intensively studied at many levels. Yet almost nothing is known
about the molecular nature of the “clocks” underlying such
rhythms. The isolation of mutations that perturb or abolish
rhythms (reviewed by Feldman, 1982) has allowed molecular analyses of these phenomena. The period (per) locus
of Drosophila melanogaster, discovered by Konopka and
Benzer (1971) is the first gene that was mutated to disrupt
biological rhythms. Different mutant alleles at this X-chromosomal locus can lengthen (per” and perLq, shorten
(per”), or abolish (per”’ and per”) circadian rhythms
(reviewed by Konopka, 1984). These mutations similarly
affect a short-term oscillation in the male’s courtship song
(Kyriacou and Hall, 1980). Therefore, this locus appears to
be centrally involved in more than one of the fly’s rhythms.
The per locus has been extensively analyzed genetically
and mapped to the 3B1-2 region of the X chromosome
(e.g., Smith and Konopka, 1981). We and others have
begun molecular analysis of the per region (Reddy et al.,
1984; Bargiello and Young, 1984). Some of our previous
findings are summarized in Figure IA, including the molecularly mapped sites of certain chromosomal lesions near
per. One of these-a
breakpoint in the T(1;4)JC43 translocation, which causes aberrant circadian rhythms (Smith
and Konopka, 1981)-appears
to be in the center of the
per region. This breakpoint also leads to lower than normal
levels of certain transcripts (Reddy et al., 1984; Bargiello
and Young, 1984) whose genomic sources are also depicted in Figure IA. We have suggested that the 0.9 kb
RNA species (see large circle in Figure IA) is of special
significance because of its abnormally low levels, not only
in flies hemizygous for the T(7;4)JC43 breakpoint but also
in both peP’ and per”’ adults (Reddy et al., 1984). In
addition, the abundance of this RNA species fluctuates
dramatically over the course of circadian cycles (Reddy et
al., 1984).
Against this background, we developed a behavioral
assay for per+ DNA in order to probe the biological activity
of defined portions of this clock locus. Two of the most
extensively characterized behavioral phenotypes known to
be affected by per mutations are circadian rhythms of
locomotor activity and rhythmic oscillations in the interpulse
intervals of the male’s courtship song. Using per” hosts
that are arrhythmic for both of these behaviors (Konopka
and Benzer, 1971; Kyriacou and Hall, 1980) we carried
out a series of P-element-mediated
transformations involving DNA obtained from the per region of wild-type flies.
We initially intended (and began experiments) to scan this
region by testing a series of adjacent and overlapping
restriction fragments
without any predisposition
as to
which segment or segments might rescue the effects of
this mutation. For reasons referred to above (cf. Reddy et
al., 1984) we decided to concentrate on DNA fragments
that include or are near the site of the T(7;4)JC43 381-2
breakpoint. We now report that either of two overlapping
DNA fragments, transduced into an arrhythmic per’ host,
rescues behaviorally mutant phenotypes. This is the first
identification, by a transformation bioassay, of a functioning part of a metazoan’s genetic locus that could not be
molecularly identified from a priori knowledge of a gene
product.
Results
Circadian
Rhythms of Transformants
Table 1 summarizes behavioral and genetic data on the
transformants. Transformation with two of the DNA fragments shown in Figure IA rescued per”‘-induced
arrhythmicity of locomotor behavior. These clones came from a
portion of the per region that crosses the T(1;4)JC43
breakpoint. One such fragment, 8.0 kb in length, is a
subset of the other rescuing DNA segment, which extends
6.6 kb farther to the right (Figure IA). The majority of the
individual flies hemizygous for per”, carrying either the
14.6 or 8.0 kb fragment, were rhythmic in their locomotor
activity. Moreover, such circadian rhythms were detected
in adults from virtually all of the independently
isolated
“14.6” or “8.0” strains (Table 1). Yet these phenotypes were
abnormal in two ways: the proportion of rhythmic cases
was relatively low, compared to the very high fraction of
wild-type flies that exhibit periodicity in their locomotor
activity (Table 1, cf. Konopka and Benzer, 1971; Smith
and Konopka, 1981); and the circadian periods for the
transformations (examples shown in Figure 2) were routinely longer than the normal 24 hr. In some cases, per”
flies carrying an 8.0 kb insert seemed to show at least two
cycles of long-period rhythmicity, which then graded into
Cdl
370
1981) a pattern of phenotypes similar to those expressed
by the “8.0” or “14.6” transformants.
D/I/IX95
1.7Kb
Adh’XbaI
Fragment
100 bp PolyLinker
ECoRI
BqllI
Figure 1. Molecular Map of Subclones in the par Region Used in Transformation, with Reference to Chromosomal Breakpoints and Transcripts from
This Region
(A) Restriction sites, cut by the endonucleases
Eco RI (R, or @ for an
artificial, nongenomrc cloning site), Barn HI (B), and Hind III (H) are shown
on the horizontal line near the bottom. The genomic sources for the 4.5,
1 .O, 0.9, 2.7, and 1.7 kb RNA species are displayed below this line. See
Reddy et al. (1984) for the relevant Northern-blot analyses, including a
determination
of the directions of transcription,
indicated here by the
arrowheads; the boundary between genomic subsegments
hybridizing to
the 2.7 versus the 1.7 kb transcript was marked by a Pst I restriction site.
The locations of chromosomal breakpoints associated with the T(7;4)JC43,
Df(l)K95,
and Df(l)w- @’ aberrations are as described in Reddy et al.
(1984) with the latter two lesions indicated here by shaded bars; horizontal
lines connected to the bars indicate DNA, as opposed to its absence,
which refers to the region to the right of the per-region breakpoint in the
Df(7)w-Hddeletion;
the Df(l)K95 lesion depicted (with lines running in both
directions from it) is an inversion breakpoint associated with this deletion
(M. Goldberg, personal communication,
and see Reddy et al., 1984). The
two Df breakpoints shown may define the boundaries of the per region,
since both of these aberrant genotypes complement per mutations (Smith
and Konopka, 1981). The left end of this map is distal on the X chromosome
(i.e., farthest from the centromere).
Solid bars indicate the positions of
fragments
conferring
rhythmicity
when transformed
into per”’ hosts,
whereas open bars indicate regions that did not transform par“’ hosts to
rhythmrcity (see Table 1).
(B) Map of the pPA-I transformation
vector (J. Posakony, unpublished),
containing the wild-type allele of the alcohol dehydrogenase
(Ad/t+) gene
of D. melanogaster and a 100 base pair polylinker located between Adh+
and P-element DNA.
arrhythmic behavior (- -> Arrh. in Table 1). This is a
phenotype similar to that expressed by certain individuals
which are hemizygous for the T(1;4jJC43 per-region breakpoint (Smith and Konopka, 1981). This same genotype
also causes stable long-period rhythmicity or complete
arrhythmicity for a given individual (Smith and Konopka,
Courtship Song Rhythms of Transformants
The 14.6 and 8.0 kb subsegments of the per region also
rescued the arrhythmic courtship singing behavior of perof
males (cf. Kyriacou and Hall, 1980). That is, most of the
males carrying either type of insert exhibited regular oscillations of the interpulse intervals (IPls) computed from
visual displays (see Figure 3A) of the acoustical output of
the wing vibrations they direct at females. The rescue of
this short-term song rhythm had a somewhat higher overall
penetrance in both sets of transformants than the circadian
phenotype (Table 1).
Six of the males (carrying 14.6 kb inserts) whose songs
were recorded and analyzed had previously been tested
for circadian rhythmic@. Three of them were rhythmic in
both locomotor and singing behavior, two were arrhythmic
for both characters, and one exhibited no circadian rhythm
but subsequently sang in a rhythmic manner.
As in the case of circadian rhythms, the song rhythm
periods were longer than normal (i.e., -70-80 set instead
of the usual per+ value of -55-60 set, cf. Kyriacou and
Hall, 1980, 1984) in perof males transformed with either a
14.6 or 8.0 kb fragment from the per region (Table 1,
Figure 3B). In one case, the entirety of the male’s -6 min
song was not rhythmic, but three strong cycles of regularly
oscillating IPls (with a period of -65 set) were computed;
the functions applied to these data (cf. Kyriacou and Hall,
1980) showed that the remaining 3 min of IPI fluctuations
was arrhythmic (- -> Arrh. in the right-hand part of Table
1). This phenotype is analogous to that exhibited by certain
of the transformed
individuals monitored for circadian
rhythms of locomotor activity (left-hand part of Table I),
as described above.
Nonrescuing
per-Region
Subsegments
Subsegments of the per region other than the 14.6 and
8.0 kb fragments did not rescue mutant behavioral phenotypes The most extensively analyzed “negative” case is
the 9.8 kb fragment (Figure IA); analysis of the relevant
transformants showed only one feebly rhythmic animal out
of 32 tested (Table 1). For the other per-region subsegments, however, there was only one transformed line (per
fragment) available for behavioral testing. Also, the courtship song has not been as extensively analyzed as locomotor activity (Table 1).
Discussion
Specific Implications
of the Transformants’
Phenotypes
The most straightforward
interpretation of our positive
behavioral results from the transformants is that the genomic change in the perO’ mutant is located within the 8.0
kb subsegment of the per region.
Since flies from the transformed lines in which the
phenotypes caused by peP
are ameliorated express
f
3
and Behavioral Analysis
26.7
27.8
27.0
:
33
0
14
__
Cl
0
0
5
27.3
26.0
2
*
*t
f
i
f
0.1
0.8
0.9
00.3
II
0.1
23.8
ca.
_-
*
_-
-_
__
_-
27
0.1
Number
Arrhythmic
Involving per-Region
28.5-->Rrrh.
ca.
2,-->Acfh.
_ca.
28
ca.
28-->Arrh.
ca.
23
26.0
27.4
3
14
of Transformants
umber
trong1y
__
24.0
14
0
__
__
__
_-
0
0
0
0
-_
__
__
__
__
25.9
8
0
0
__
__
_-
24.8
3
+
+
+
__
+
+
f
+
+
f
__
_-
27.0
+
25.0
26.7
25.3
27.3
26.8
26.5
27.6
0.1
0.6
0.6
0.2
0.3
0.4
0.2
1.3
0.4
1.0
Period
(h f SE")
man
1"
29
7
6
4
5
7
__
hythmic
II
DNA
-
8
”
R
0
0
0
1
__
0
1
__
__
7
__
__
5
0
2
2
1
__
:
.c
24.3
ca.
ca.
25.1
-_
+
__
__
--
__
__
__
__
f
f
__
__
0.3
0.6
1.3
0
0.4
0.2
22
22
25.5
+
27.0
__
23.3
f+
30.3
Period
lh * SEMI
24.9
Locomotor Activity
(periodogramsl
3
10
0
12
8
8
:
__
24
:
-_
-_
--
__
+
__
_-
68.2
__
0
7
7.3
--
0
*
2
0
3
1
--
__
__
*
__
1
1
__
--
1
Y
__
__
0
--
0
0
__
The transformed strains are named (leftmost column) with respect to the lengths In krlobases of per-region subsegments
(see Frgure IA) used to produce me vectors that allowed generation of the transformants,
Each strain also is named by the addition of an arbitrary number to desrgnate it as independently isolated line (i.e., more than one strain was usually isolated from the series of injections involving a given per-region
fragment). Orientations of per-region inserts (in the transformation vectors and in the genome of transformants)
are indicated in the first column as follows: for the 14.8, 7.3, and 8.0 kb inserts, (a) indicates the 0.9 kb
RNA is transcribed antiparallel to the Adh+ RNA, and (b) indicates the 0.9 kb RNA is parallel to the Adh+ RNA; for the 9.8 kb Inserts, (a) indicates the 2.7 kb RNA is antiparallel to the Adh+ RNA; for the 5.8 kb insert
(b) indicates the 1.7 kb RNA is parallel to the Adh+ RNA; for the 9.0 kb insert (a) indicates the 4.5 kb RNA is antiparallel to the Adh+ RNA (see Figure 1B and Experimental Procedures),
The genomic locations of the insert (or inserts) for each strain are listed in the second column; some transformants
carry at least one insert on two or more separate chromosomes.
Data from many of the activity tests were collected as actograms (see Figure 2) and assessed for rhythmicity versus arrhythmicity as in Konopka and Benzer (1971). The periods were estimated as described rn
Expenmental Procedures. The results of the periodogram analyses of circadian rhythmicity (see Experimental Procedures) were divided into three categories: strongly rhythmic (a robust and wall defined peak in the
perk&gram,
corresponding
to a given period which was determined to the nearest half-hour); weakly rhythmic (a relatively small peak, due to a somewhat sloppy rhythm, a rather inactive fly, or both); and arrhythmic
(no discernible peak).
The overall rhythmicity of the “8.0” strains was stronger, as a result of the periodogram analyses (see Experimental Procedures),
than from the appearance of the actograms. Yet an inspection of the digital data
(number of activity events in SUCCSSSive
time frames), to which the periodogram functions were applied in analyzing these strains, led to the somewhat subjective impression that these flies expressed weaker circadian
rhythms (digitally determined) than in the wild type. That is, the overall pattern of results on the “8.0” transformants
seemed, from inspection of both kinds of relevant data (analog or digital), to be similar,
In the song tests, rhythmicity (or lack of it) in the fluctuations of pulse song IPls was determined as described in Kyriacou and Hall (1980) summarized here in Experimental Procedures and the legend to Figure 3.
9.8:9(a)
9.8:10(a)
9.8:16(a)
9.8:26(a)
9.8:*9(a)
Table 1. Genetic
Cell
372
perti; P[ per’] Adhf”23pr cn
.-
14.6:21
I
.
I
~~26.5 h
~=27 h
...z... -.
perd;P[per’]
Adhf”23pr cn
14.6:63
.
I-.
~=26 h
rx23.7
-_-
h
per”
Figure 2. Actograms
of Locomotor Activity
The genotypes of the transformants
shown (as opposed to the per + and per” controls) are according to the nomenclature prescribed in Spradling and Rubin
(1983). In the actograms time moves from left to right and from top to bottom (see Expenmental Procedures). The data displayed represent recordings made
during freerunning (Le., after entrainment of the adults by light/dark cycles). The black bar in the center of each set of traces corresponds
to the time when
the lights had been turned on during each of the three entrainment days.
The normal (per’) fly had activrty offsets and onsets, which are arranged rn an approximately vertical manner (slope -0, see Experimental Procedures),
near the beginning and the end (respectively)
of the subjective day (hence 7 = -24 hr). However, the activity offsets for the transformants move progressively
to the right on successive
days; these shifting offsets, which are later and later than in the normal case, therefore define a positively sloping line (see
Experimental Procedures), and this is reflected in the relatively long periods (8s) depicted here for the transformants
from different “14.6” strains (also see
Table 1)
either long-period rhythms or are arrhythmic, we infer that
the expression of per-region functions is subnormal in
these two types of transformants. This conclusion is based
on the supposition that long-period perL variants seem to
be hypomorphic
mutations (summarized
in Konopka,
1984) and that hemizygosity for the per-region breakpoint
in 7(7;4)JC43 leads to an overall array of locomotor activity
phenotypes not unlike those exhibited by our transformants. The chromosomally aberrant genotype, moreover,
causes hypomorphic expression of three transcripts from
the per region (Reddy et al., 1984).
Hypomorphic expression of per can also be achieved
simply by constructing females that are heterozygous for
peP’, or a deletion of the locus, and per+. In these cases,
circadian periods are slightly longer than normal (-25 hr),
but essentially every individual is rhythmic (Konopka and
Benzer, 1971; Smith and Konopka, 1981). We infer that
the array of behavioral phenotypes expressed by our
transformants is not due merely to the fact that the various
transformed flies carried only one dose of transduced per+
DNA (also, some of the flies we tested carried more than
one dose, but were still long-period; see Experimental
Procedures).
We do not have enough molecular information to interpret the apparently subnormal activity of the two rescuing
DNA segments in the transformants. One possibility is that
sequences adjacent to these pieces of DNA are necessary
for completely normal per+ expression. It is also conceivable that the various genomic sites at which the 8.0 and
14.6 kb fragments from the per region have integrated
;raFformation
of per
IPI
A
lead to subnormal expression; this would include the possibility that autosomal locations for the inserts (i.e., the
great majority of the cases listed in Table 1) are somehow
associated with relatively poor transcription of the relevant
RNAs, which are normally transcribed from the X chromosome. It is known, however, that ectopic locations for
autosomal enzyme-encoding
loci in D. melanogaster usually (Scholnick et al., 1983; Spradling and Rubin, 1983)
but not in every case (Goldberg et al., 1983) allow for
nearly normal expression of these transduced genes; and
that autosomal DNA inserts of the (usually X-chromosomal)
white (w) locus typically-but,
again, not always-give
apparently normal levels of w+ expression (Hazelrigg et al.,
1984).
The failure of a given fragment to affect the behavior of
the per0 hosts must be interpreted with caution. For example, the 9.8 kb fragment fails to rescue the per“’
mutation; thus the 2.7 kb RNA species (Figure 1A) would
seem to be eliminated as a candidate for the transcript
that is directly affected by per” mutations. Yet the various
RNAs transcribed from this region have not been mapped
with enough precision to show whether an individual segment of the per region contains all of the coding (and
noncoding)
information required to express properly a
given transcript. Experiments to assess the capacity of the
various transduced DNA fragments to express, in vivo,
individual transcripts are in progress.
The results of this study, taken together with our previous
findings (Reddy et al., 1984) suggest that either the 0.9
kb transcript or the 1.O kb RNA species (see Figure 1A) is
the transcript directly affected by peP’, and that this
mutational defect is rescued in Pans by introduction into
the genome of wild-type DNA homologous to either or
both of these transcripts. We propose further that the 1 .O
kb transcript may be the key RNA species; this is because
the 7.3 kb DNA fragment-which
looks as if it should be
able to produce the 0.9 kb RNA but not the 1 .O kb RNA
(see Figure lA)-did
not rescue per”‘-induced
arrhythmicity (although only one such “7.3” strain was recovered,
Table 1). This speculation is correlated with, and influenced
by, previous genetic and behavioral data implying that Xchromosomal material to the left of the T(l;4)JC43 breakpoint (to which the 1.0 but not the 0.9 kb transcript is
homologous)
is sufficient to confer weak, long-period
rhythmicity in circadian behavior (Smith and Konopka,
1981). If the 1 .O kb RNA is the species affected directly
by peP’ and in need of rescue, our present results and
those reported earlier (Reddy et al., 1984) imply that the
100 ms
B
Per “‘; P[pw+]
Adbfnz3prcn
14.6~63
r=66.1
s
46r
per”
P[perf];Adhfn”Prcn
per”‘;
P [per +
8.0~2
Adhfnz3prcn;
P[perf]
r=63.1
s
9.8:lO
cl
48
0
0
36
Figure 3. Oscillatrons
of Interpulse
min
Intervals in the Male’s Courtshrp
Song
Males of the indicated transformed genotypes (cf. Table 1 and Figure 2)
had their songs recorded and analyzed as described in Experimental
Procedures. (A) An example of a pulse train, with eight tone pulses, hence
seven interpulse intervals (IPls), is depicted. In this species of Drosophila,
pulses are typically generated at the rate of -20-30/set
(e.g., Kyriacou
and Hall, 1980).
(6) The two rhythmic songs (for a male from a “14.6” transformed line, and
one from an “8.0” line) are indicated by the sinusoidal (dashed) lines,
accompanied
by the best estimates of the periods (T’S). No line was drawn
with respect to the IPI fluctuations for the male from a “9.8” transformed
lrne (cf. Table I), because there was no discernible pattern to these
changing IPls.
Cell
374
dramatic effects of arrhythmic per mutations on the abundance of the 0.9 kb transcript are not the primary cause
of arrhythmicity.
More complex models are possible and cannot be
excluded at present. Among these is the possibility that
the per”’ mutation does not map within the 8.0 kb subsegment of the per-region but has an adverse effect in cis on
a transcript or transcripts coded for by this subsegment
of the per region. Such a model implies that one or more
transcripts coded for by this fragment (presumably the 1 .O
and/or the 0.9 kb RNA species) are able to provide for
rhythmic behaviors when the 8.0 kb segment is not adjacent to the per"' mutation. Therefore, the site of this
mutation would not necessarily define a portion of the per
locus that, in the wild-type, is required for downstream
regulation of the transcription involved in the control of
normal rhythmicity. This possibility (and others) can be
examined by molecularly defining the site of the peror
mutation.
General Implications
of the Molecular
Expression
of the per Locus
As for other biological rhythms in Drosophila-and
their
regulation by expression of the per locus-no
single transcript seems sufficient to control all the relevant phenotypes For instance, erratic heartbeating has been detected
in perO’ larvae (E. 0. Aceves-Piha and M. S. Livingstone,
unpublished;
see Livingstone,
1981). The RNAs transcribed from the portions of the per region immediately
adjacent to the T(1;4)JC43 breakpoint (Figure 1A) would
appear to contain inadequate information to influence this
phenotype, because none of these transcripts is detectable earlier than the late pupal stage of development
(Reddy et al., 1984; Bargiello and Young, 1984). Two of
the other per-region transcripts shown in Figure IA, the
2.7 and 1.7 kb species, are present during all stages of
the life cycle, though only the former RNA is detectable in
the fly’s head (see Reddy et al., 1984, for discussion of
this and other evidence suggesting that the 1.7 kb transcript may be associated with the expression of a separate
gene, located to the right of per, called /(l)zw6). Therefore,
the 2.7 kb transcript is the only obvious candidate for
participating in the control of rhythmic phenotypes that are
entrained or expressed relatively early in development and
are also disturbed by per mutations (see further discussion
of these several phenotypes in Konopka, 1984).
Another rhythmic phenomenon associated with the per
locus, the light/dark-mediated
fluctuation in the 0.9 kb
transcript’s abundance (Reddy et al., 1984) is almost
certainly related to the fly’s circadian rhythms of locomotor
activity. However, the correlation of this molecular phenotype with the male’s song rhythms is ostensibly obscure,
since that oscillating behavior is insensitive to changing
environmental cues (Kyriacou and Hall, 1980). For these
reasons, and because of the molecular complexity of the
per region’s transcripts, we suspect this locus may contribute more than one functional product to the control of
the fly’s biological clocks, Consequently, transformation
experiments must be extended to an analysis of various
other per-mutant phenotypes, such as eclosion and the
larval heartbeat. Also, additional transformed lines must be
generated both by injection of as yet unanalyzed portions
of the per region and by genetic and molecular construction of various combinations of DNA fragments. The results
presented here represent an important first step in this
extensive program and lead to the strong prediction that
transcripts (and proteins?) coded for by the 8.0 kb subset
of the per locus are components of the control of Drosophila’s biological clocks.
Experimental
Procedures
Construction
of Transformation
Vectors
In preparation for generation of the transformants,
overlapping Eco RI or
Barn HI restriction fragments were ligated to the P-element transformation
vector pPA-1 (Figure 1 B, J. Posakony, unpublished) that had been linearized
by digestion with either Eco RI or Barn HI. Ligated DNA was transformed
into the DHl strain of E. coli. Plasmid DNA was extracted from ampicillinresistant colonies and analyzed by restriction digestion as described by
Reddy et al. (1984). Restriction endonucleases
and T4 DNA polymsrase
were purchased from New England Biolabs and Bethesda Research Labs
and used as recommended
by the manufacturers.
By using cuts made in
the polylinker of the transformation
vector (Figure IB), clones containing
the fragments indicated in Figure IA were inserted in the two possible
orientations with respect to the Adh gene; the orientations were confirmed
by analysis of restriction digests,
Generation
of Transformants
The transformants
were produced by microinjection of cloned DNA (cf.
Spradling and Rubin, 1982; Rubin and Spradling, 1982). We used a series
of transformation
vectors in which the separate DNA fragments from the
per region (Figure 1A) had been cloned into the plasmid shown in Figure
1B. All DNA transformation cocktails, with the exception of that involving
the 5.8 kb fragment, also contained the helper P-element plasmid
p?r25.7wc, which is able to provide “transposase” function but is unable to
insert into the D. melanogaster genome (Karess and Rubin, 1984). The
“5.8” transformations
utilized the helper P-element plasmid pr25.1, able
both to provide transposase function and to be inserted into a chromosome.
For both types of helper plasmid, this DNA was present at a concentration
of 50 ag/ml. In the cases of transformations
involving the 5.8, 7.3, and 9.0
kb fragments (Figure IA), roughly equal amounts of the per-recombinant
plasmids, containing single inserts in both possible orientations, were
present in the transformation cocktails at a final concentration of 500 pg/
ml. The 9.8 and 8.0 kb fragments were cloned into pPA-I (Figure 1B) in
each orientation, but injections were carried out separately with each
orientation of the insert, The 14.6 kb fragment was inserted in only one
orientation, with the 0.9 kb RNA antiparallel to the Adh+ RNA. In all these
cases, where a single orientation of the fragment in the vector was injected,
the DNA was again at 500 pg/ml. Each transformant line was examined by
Southern blotting (data not shown) to determine the orientation of the
inserted DNA wrth respect to the Aclh+ gene present in the vector (see
Table 1).
Microinjection of transforming DNA was carried out using the techniques
of Rubin and Spradling (1982) as applied by Goldberg et al. (1983). The
latter authors rescued
the effects of an alcohol-dehydrogenase-null
(Adh’“23) mutation (cf. Benyajati et al., 1983). Our injections were performed
on embryos homozygous for this mutation (linked to the eye color markers
pr and cn); in most cases, the injectees were also homozygous
or
hemizygous for per”‘. Adults were collected as they emerged rn the cultures
that resulted from the development
of the injected embryos. Such flies
were mated srngly to pep’; Adh’“23 pr cn (or, when appropriate, Adh”‘23
pr cn) adults to produce a series of strains in the next generation (called
Gl). These Gl strains were screened for acquired resistance to a 6%
ethanol-water
solution in a test descrrbed by Vigue and Sofer (1974). It
was important to age the adults for 4 days before ethanol treatment (to
avord the misleadrng sun4val of nontransformed
Adh”‘= individuals).
Transformation
375
of per
Basic Characterization
and Manipulation
of Transformants
Strains recognized to contain Adh+ inserts have been maintained by regular
selection for ethanol resistance, but were behaviorally tested soon after
therr isolation. The chromosomal location of each insert was determined by
applying standard genetic techniques for assessment
of X linkage versus
autosomal inheritance. In the latter cases, crosses involving autosomaldominant-containing
balancer chromosomes
led to an assignment of 2nd.
versus 3rd.chromosomal
linkage of the insert(s) for a given strain, using
ethanol selection and spectrophotometric
assays (re ADH activity, cf.
Ursprung et al., 1970) of the appropriate segregants.
For the transformants
carrying the 5.8, 7.3, and 9.0 kb fragments, which
were initially obtained in a per’ background,
it was necessary to cross
individual ethanol-resistant males to per”‘; Adh “‘23prcn females and select
ethanol-resistant
male progeny (hemizygous for per”) prior to behavioral
testing. All other transformations
were carried out in the per”‘; Adh”‘*3 pr
cn genetic background.
For all the transformed strains, the presence of the
insert from the per region was confirmed (data not shown) by Southern
blot analysis, using appropriate probes (Reddy et al., 1984) and techniques
similar to those previously reported (e.g. Goldberg et al., 1983; Hazelrigg
et al., 1984).
After selecting individual Adh+ males from a given strain containing an
insert from the per region, these flies were monitored for circadian rhythms
for locomotor activity or were recorded with respect to courtship singing
behavior, as described below.
Analysis
of Circadian
Rhythms
Single male flies were selected (if appropriate, see Table 1) and exposed
to 3 days of entrainment in light/dark cycles of 12 hr/12 hr. At the end of
the third entrainment day, a given transformed or control fly was marntained
for 4-7 additional days in constant darkness,
when the “freerunning”
endogenous rhythms (if any) of locomotor activity are expressed. For each
Individual fly, activity was monitored by methods similar to those previously
reported (Konopka and Benzer, 1971; Smith and Konopka, 1981; Jackson,
1983). An infrared light beam was passed through the short axis of a
cylindrical glass tube (15 mm long X 4 mm diameter), and a photoelectric
cell picked up any interruptions in the light beam caused by the fly moving
Into it. These momentary signals were amplified and recorded on chart
paper as a function of time. In addition, for many of the data processed
and then summarized in Table 1, activity events were recorded digitally as
described in Smith and Konopka (1981) and Sulzman (1982).
Subsequent to collection of data by the event recorders, markings on
the chart paper were analyzed by cutting strips corresponding
to 24 hr
segments and placing each subsequent strip in a vertical column (cf. Figure
2). During entrainment, per”’ (as well as wild-type) adults are rather rhythmic
in their activity (data not shown in Figure 2) in that the flies are simply more
active rn response to light (hence these behavioral fluctuations are not,
strictly speaking, endogenous rhythms).
The period length (in hours) for a given fly was defined as 7 = 24 + s,
where s is the slope (in hours per day) of a line drawn so that it passes
approximately through the activity “offsets” during freerunning. An offset is
the termination of densely recorded markings within a given rest/activity
cycle. By convention, the lines were drawn through these offsets with the
chart paper strips oriented vertically (so that offsets occurring about the
same time in each 24 hr time frame define azero slope, and those occurring
later and later in successive
cycles define a positive slope; cf. Figure 2).
Most of the flies tested behaviorally were not known to be heterozygous
or homozygous for an insert or inserts. Yet, in some cases, individuals were
subsequently
outcrossed
to Adhmz3 pr cn flies, and the Fi tested for
ethanol resistance.
Two such crosses (involving “14.6” transformants)
revealed that the behaviorally tested fly had been homozygous for an insert
(or inserts); yet the behavioral phenotypes (longer than normal periods, i.e.,
both 27 hr) were similar to those determined for other “14.6” individuals,
whose progeny revealed that they had carned only one “dose” of an Adh+containing autosome (mean T, for 17 rhythmrc flies, 26.9 f 0.2 hr; arrhythmic, n = 9).
In conjunction
with tests of circadian rhythmrcity
of ethanol-treated
transformants,
the same treatments of per+; Adh+ adults (from a CantonS strain, which was the source of normal controls and from which the
original per mutations were derived, Konopka and Benrer, 1971) showed
that the treated flies were rhythmic in either behavioral character, with no
period lengthening (cf. such effects, in other organrsms, when they are
monitored for circadian rhythms rn the chronic presence of alcohol; eg,
Biinnrng and Moser, 1973). Additional controls showed that the Adh’“23 pr
cn chromosome
(when homozygous
or heterozygous
with a Canton-Sderived 2nd chromosome)
had no effect on normal circadian rhythmicity
(in a per+ background)
or on the arrhythmicity
induced by per”‘. A few
control transformants
were obtained in which the hybrid P-element/Adh+
vector (see Figure lB), with no per-region insert, was injected: individuals
from the resultant ethanol-resistant
strains (when hemizygous for per”‘)
tested as arrhythmic (data not shown).
The activity data collected digitally (see above) were analyzed using
periodogram functions (cf. Smith and Konopka, 1981; Sulzman, 1982) to
determine objectively whether the activity of a given fly was rhythmic and
to obtain the best estimate of the circadian period. In the preliminary tests
of this type, the data were collected and the periodograms generated with
no knowledge of whether the transformed
strains in question appeared,
from the actogram analyses, to contain rhythmic flies (Le., these tests of
the digital data were done “blrnd”).
Analysis
of Courtship
Song Rhythms
Before being tested in courtship, males were aged, ethanol treated (see
above), and in some cases monrtored for circadian rhythms. Subsequent
to these treatments and tests, males were maintained singly for one more
day in unyeasted food vials, and then each was paired with a single, young,
virgin female from a wild-type (Canton-S) stock. Courtship typically began
shortly after Introduction
of the two flies into the recording chamber,
consistrng of a small, nylon-mesh-covered
plastic cell that was placed close
to an electret microphone; thus device had been modrfred to be sensitrve
to the “parkcle velocrty” component of sound waves (see Bennet-Clark,
1984, for brologrcal and electronic details). Sounds produced by the courting
male’s wrng vibratrons were amplified and recorded (on magnetic tape) for
-4-7 mm of courtship. The signals were then transferred to oscillograph
paper, on which pulse trains appeared as exemplrfred in Frgure 3A. Analysrs
of the successrve
Intervals between pulses (IPls) was carried out by the
method of Kyrracou and Hall (1980) in which a mean for all the IPls
occurrrng within a given 10 set time frame IS computed. The means for
successrve frames were plotted (as shown rn Figure 38) wrth the provrso
that at least 15 pulses were generated by a male srnging durrng that
particular 10 set period. Srnusordal curves (dashed lines In Figure 38) with
the best-fitting periods and amplrtudes, were drawn only if a nonlinear
regressron analysis and accompanying
F-ratio statistic revealed that the IPI
fluctuations were significantly rhythmic at the 5% level or beyond. Thus krnd
of analysis typically reveals that most, but not all, fires expressing per+ sing
with regularly oscillating IPls (Kynacou and Hall, 1980, 1984). All of the
present data collectrons and analyses (for transformants and controls) were
performed “blrnd,” such that the genotype associated with each song was
decoded only after a formal assessment of rhythmicity versus arrhythmicity.
Acknowledgments
We thank J. Posakony for the P-element/Adh+
vector, the ~~25.1 plasmid,
and the Adhfiz3 stock of Drosophrla melanogaster. We are grateful to G.
Rubin for the pa257wc
plasmid. We greatly appreciate electronic aid from
D. Wight, H. Bennet-Clark, and M. Gorczyca and are extremely thankful for
the data acquisition and analysis software donated by F. Sulzman. We
thank J. Posakony, A. James, and M. Gray for comments on the manuscript.
This work was supported by National Institutes of Health grant GM-33205
to M. R. and J. C. H., by a Clarkson College grant to R. J. K., and by SERC
grant 94595 to C. P. K.; W. A. Z. is supported by a National Institutes of
Health postdoctoral fellowship, and P. R. by a Dretzin predoctoral fellowship.
The costs of publication of this article were defrayed in part by the
payment of page charges. This article must therefore be hereby marked
“advertisement”
in accordance
with 18 U.S.C. Section 1734 solely to
indicate this fact.
Recerved
September
7, 1984
References
Bargrello, T. A., and Young, M. W. (1984). Molecular genetics of a biological
clock in Drosophila. Proc. Nat. Acad. Sci. USA 81, 2142-2146.
Cell
376
Bennet-Clark,
small Insects
463.
J. C. (1984). A particle velocity microphone for the song of
and other acoustic measurements.
J. Exp. Biol. 108, 459-
Benyajati, C., Place, A. R., Wangh, N., Pent& E., and Sofer, W. (1983).
Formaldehyde
mutagenesis
in Drosophila:
molecular analysis of ADHnegative mutants. Mut. Res. 7 17, 1-7.
Bunning, E., and Moser, I. (1973). Light-induced phase shifts of circadian
leaf movements of Phasaohx
comparison with the effects of potassium
and of ethyl alcohol. Proc. Nat. Acad. Sci. USA 70, 3387-3389.
Feldman, J. F. (1982). Genetic
Plant Physiol. 33, 583-608.
approaches
to circadian
clocks.
Ann. Rev.
Goldberg, D. A., Posakony, J. W., and Maniatis, T. (1983). Correct developmental expression of a cloned alcohol dehydrogenase
gene transduced
into the Drosophila germ line. Cell 34, 59-73.
Hazelrigg, T., Levis, R., and Rubin, G. M. (1984). Transformation
of white
locus DNA in Drosophila: dosage compensation,
zeste interaction, and
position effects. Cell 36, 469-481.
Jackson, F. R. (1983). The isolation of biological rhythm mutations
autosomes of Drosophila melanogasfer. J. Neurogenet. 7, 3-15.
Karess, R. E., and Rubin, G. M. (1984). Analysis
functions in Drosophila. Cell 38, 135-146.
Konopka, R. J. (1984). Neurogenetics
Evolutionary Genetics of Invertebrate
York: Plenum Press), in press.
of P transposable
on the
element
of Drosophila circadian rhythms. In
Behavior, M. D. Huettel, ed. (New
Konopka, R. J., and Benzer, S. (1971). Clock mutants
melanogaster. Proc. Nat. Acad. Sci. USA 68, 2112-2116.
of Drosophila
Kyriacou, C. P., and Hall, J. C. (1980). Circadian rhythm mutations in
Drosophila melanogasfer
affect a short-term fluctuation in the male’s
courtship song. Proc. Nat. Acad. Sci. USA 77, 6929-6933.
Kyriacou, C. P., and Hall, J. C. (1984). Action potential mutations
biological clock in Drosophila. Nature, in press.
stop a
Livingstone, M. S. (1981). Two mutations in Drosophila affect the synthesis
of octopamine, dopamine, and serotonin by altering the activities of two
amino-acid decarboxylases.
Neurosci. Abstr. 7, 351.
Reddy, P., Zehring, W. A., Wheeler, D. A., Pirrotta, V., Hadfield, C., Hall, J.
C., and Rosbash, M. (1984). Molecular analysis of the period locus in
Drosophila melanogaster and identification of a transcript involved in biological rhythms. Cell 38, 701-710.
Rubin, G. M., and Spradling, A. C. (1982). Genetic transformation
Drosophila with transposable element vectors. Science 278, 348-353.
of
Scholnick, S. B., Morgan, B. A., and Hirsh, J. (1983). The cloned dopa
decarboxylase
gene is developmentally
regulated when reintegrated into
the Drosophila genome. Ceil 34, 37-45.
Smith, R. F., and Konopka, R. J. (1981). Circadian clock phenotypes of
chromosomal
aberrations with a breakpoint at the per locus. Mol. Gen.
Genet. 783, 243-251.
Spradling, A. C., and Rubin, G. M. (1982). Transposition
of cloned P
elements into Drosophila germ line chromosomes.
Science 278, 341-347.
Spradling, A. C., and Rubin, G. M. (1983). The effect of chromosomal
position on the expression of the Drosophila xanthine dehydrogenase
gene.
Cell 34, 47-57.
Sulzman, F. M. (1982). Microcomputer
Comput. Biol. Med. 72. 253-261.
monitoring
of circadian
rhythms,
Ursprung, H., Sofer, W., and Burroughs, N. (1970). Ontogeny and tissue
distribution of alcohol dehydrogenase
in Drosophila melanogaster. Wilhelm
Roux’s Arch. 164, 201-208.
Vigue, C., and Sofer, W. (1974). Ad/?“? a temperature-sensitive
the Adh locus in Drosophila. Biochem. Genet. 74, 127-135.
mutant at