-
The EMBO Journal vol. 1 3 no. 7 pp. 1 636 - 1644, 1994
The basis for germline specificity of the hobotransposable
element in Drosophila melanogaster
Brian R.Calvil and William M.Gelbart2Department of Cellular and
Developmental Biology, HarvardUniversity, Cambridge, MA 02138,
USA'Present address: Carnegie Institution of Washington,
Baltimore,MD 21210, USA2Corresponding authorCommunicated by
F.Kafatos
This paper is dedicated to the memory of Barbara
McClintockPrevious results suggested that the hobo
transposableelement is active predominantly in the germline
ofDrosophila. We investigate germline restriction of
hobotransposition by testing in vitro modified elements fortheir
ability to mobilize marked elements in vivo.Although intact hobo
elements are germline specific, anhsp70 promoter-hobo transposase
fusion is active in thesoma. Analysis of the hsp7O-promoted
transcript does notprovide evidence for splicing. Moreover, the
hobopromoter confers germilne bias to a highly sensitivereporter,
A2-3 P transposase. These results indicate thathobo transposition
is germline specific due to regulationof transposase production at
the level of transcription.Thus, although hobo is similar to the P
transposableelement in organization and tissue specificity, it
differsin the underlying mechanism governing germline
specificactivity.Key words: Drosophilalgermline/hobolP
element/transposon
IntroductionIt is over 40 years since Barbara McClintock
describedgenetic instability in maize due to the action of the
Activator(Ac) transposable element (McClintock, 1948, 1952).
Sincethat time transposable elements have been found in
virtuallyevery organism studied. Genetic and molecular
investigationsof transposable elements have reformed thinking
aboutgenomic organization and dynamics, and have revealed avariety
of insights into the possibilities for
biologicalregulation.Although over-replication relative to the host
is critical
to a transposable element's persistence and spread, in mostcases
controls on replication exist to reduce deleteriouseffects on the
host [for reviews see Kleckner (1990) andSmith and Corces (1991)].
Most transposable elements haveone or more mechanisms to dampen
their level of activity.These regulatory mechanisms can act by
limiting functionaltransposase or by acting on the transposition
reaction. Forelements within multicellular organisms, one aspect of
thiscontrol can involve tissue specific transposition. Here
weexamine the regulation of the hobo transposable element
ofDrosophila, and compare the basis for its germline
specificactivity with that of the related P transposable
element.The hobo and P transposable elements in Drosophila both
belong to the Ac family of elements (O'Hare and Rubin,1636
1983; Streck et al., 1986; Calvi et al., 1991). Both P andhobo
are - 3 kb in length, exist in full-length and internallydeleted
forms, have short terminal inverted repeats of limitedsequence
similarity between the two elements, produce atrans-acting
transposase that is thought to catalyze a DNA-mediated
transposition reaction (Kaufman and Rio, 1992),and create an 8 bp
duplication of host DNA upon insertion.The transposases of the two
elements, however, do not haveamino acid sequence similarity
(Streck et al., 1986; Calvietal., 1991).The P element is among the
most intensively investigated
of all eukaryotic transposable elements [for reviews seeEngels
(1989) and Rio (1991)]. P activity is restricted tothe germline due
to regulated RNA splicing (Laski et al.,1986; Rio et al., 1986). In
germline cells three introns arespliced out of the P transcript
leading to the production oftransposase. In contrast, in the soma
the third intron is notremoved, and, due to the presence of a stop
codon withinthis intron, a truncated protein without transposase
activityis produced. This differential splicing is due to
somaticfactors that inhibit splicing of the P third intron (Laski
andRubin, 1989; Siebel and Rio, 1990; Siebel et al., 1992).Our
earlier work characterizing hobo-mediated genetic
instability at the decapentaplegic (dpp) locus revealed
anadditional similarity with P in that hobo activity appears tobe
largely restricted to the germline [for reviews seeBlackman et al.
(1987) and Blackman and Gelbart (1989)].However, sporadic
observations from our laboratory andothers suggest that in some
rare cases hobo may be activein the soma at very low frequency
(Lim, 1981, 1988;Yannopoulos et al., 1983, 1987). Interpretation of
theseresults was hampered because these earlier studies relied
onstrains containing numerous, unmarked hobo elements. Inthis
report, we use marked, nonautonomous target elementsand in vitro
modified sources of hobo transposase to examinegermline specificity
of hobo activity. Our results indicatethat, like P, hobo activity
is restricted to the germline bylimitation of transposase to this
tissue. We present evidencethat, unlike the P element, germline
specificity is due toregulation of hobo transposase production at
the level oftranscription.
ResultsHFL1 is a 3 kb hobo clone derived from the dppdblk
strainwhich contains hobo germline activity but little if any
somaticactivity (Blackman et al., 1987). The ability of HFL1
tocatalyze germline transformation in injection assays
indicatesthat it encodes hobo transposase (Blackman et al.,
1989).Our sequence analysis of HFL1 revealed a 1.9 kb openreading
frame (ORF1) that comprises the majority of hobo[base pairs (bp)
307-2289], and is similar in predictedamino acid sequence to the
transposases of the Activator andTam3 transposable elements. The
numbering of the hobosequence follows that analysis with + 1
corresponding to the
© Oxford University Press
-
hobo germline specificity
base pair of the element. We showed that, althoughsequences 3'
of ORFI are dispensable, frameshifts withinORFI inactivate
transposase (Calvi et al., 1991) (Figure 1).The only other ORF that
may be required for transposaseproduction is the upstream, 96 bp
long ORFO (bp 208 -303)which is separated from ORFI by a single
in-frame stopcodon. ORFO contains the first potential initiator
codon morethan 30 bp 3' of the hobo TATA box consensus which
residesat position 107 of HFL1. 60 bp 5' of the TATA consensus,hobo
contains a sequence that resembles the CAAT boxconsensus. There is
no further evidence, however, that thisregion contains the hobo
promoter.
Tissue specificity of hobo element activityTo overcome the
problems of multiple copy number andinstability inherent in
studying natural transposable elements,we have developed a system
whereby hobo transposasesources, rendered immobile by removal of
their terminirequired in cis for transposition, are introduced back
intothe genome via P element transformation (Calvi et al.,1991).
The subsequent transformants are genetically marked
IS-} 11H
I."
Fig. 1. Schematic representation of sources of h(their activity
in the germline and soma. For clarihobo segments are shown (see
Materials and meidescription). All derivatives are based upon the
thobo clone HFL1 (Blackman et al., 1989; Calviis shown on top with
nucleotide positions of releHBL1 is identical to HFL1 except for a
106 bpwhich does not affect transposase production butstable in the
presence of hobo transposase (Calviis a fusion of the
heat-inducible hsp70 promoter(UT) leader sequences (-245 to +207 of
hsp70site of transcription) to position 147 of hobo, 40hobo TATA
consensus, thereby retaining ORFOHBL1 and HSH2 contain hobo 3' UT
sequence zsignals. The nucleotide coordinates for the
endposequences present are shown above the elements.as a striped
box, and ORFI as a shaded box. Arboxes represent the terminal
inverted repeats.
and stable, and can be tested for hobo transposase productionby
their ability to mobilize marked hobo elements in trans.The assay
we employ here relies on mobilization of
H[w+, hawl] (hawl), a hobo element marked with themini-white
gene. hawl does not produce transposase, butis mobilized if
provided with the enzyme in trans (non-autonomous). Expression of
mini-white, a derivative of thewhite gene, within hawl is sensitive
to genomic position(Pirrotta et al., 1985). Different genomic
insertion sites ofhawl impart different levels of pigment to the
eye. Fliescontaining a null mutation at the white locus and a
giveninsertion of haw have anywhere from faint yellow to darkred
eyes. As white expression is cell autonomous, somatictransposition
or excision of hawl in the soma duringdevelopment can result in
mosaic eye pigmentation. To detecthobo activity in the male
germline, we monitor transpositionof haw 1 from the X chromosome to
the autosomes, detectedas father to son transmission of the
initially X-linked mini-white marker gene (Calvi et al., 1991;
Materials andmethods).
Germline specificity of an intact hobo transposasesourceAs an
essentially wild-type source of hobo transposase, we
St'1in.1 ( i'>r rnIiis have constructed a derivative of HFL1,
the element P[ry+,HBLl] (P-Hobbled or P-HBL1) (Figure 1). This
element
+ contains a small deletion of the 3' end of hobo, renderingit
stable but leaving transposase coding regions intact (Calviet al.,
1991). P-HBLl displayed transposase activity onlyin the germline.
When present with two copies of hawl on
- + the X chromosome, different insertions of P-HBLl
catalyzedone or more transpositions in 15-18% of the male
germlinestested (Table I). P-HBLl was also active in female
germline
+ + (unpublished data). In contrast, independent of
genomicposition, P-HBLl was not active in somatically derived
eye
obo transposase and tissue as indicated by the absence of mosaic
eye pigmentationity, only hsp70 and in flies containing the two
hawl elements on the Xthods for a complete chromosome and any of
three independent insertions of P-
et al., 1991), which HBL1 (Table I). P-HBLI was also germiine
specific whenvant landmarks. P- tested for mobilization of hawl
from other genomic positions3' terminal deletion (unpublished
data).renders the element Given the similarities in organization
and tissue specificityet al., 1991). HSH2 between hobo and P, we
first wanted to investigate theand 5' untranslated possibility that
hobo activity is restricted to the germline bybp downstream of the
regulated RNA splicing. However, Northern analysis, cDNAand ORF1.
Both P- library screening, PCR cDNA production, and RNaseand
polyadenylation protection of strains containing P-HBLI or
naturallyORFO iS represented occurring hobo elements failed to
detect hobo RNA, and thusrowheads within the were uninformative
about hobo transcript structure and the
question of splicing regulation.
Table I. Assay of P-HBL1 activity in the soma and germline
Transposase source Soma Germline
Adults Mosaic Germlines Germlines withscored adults tested
transpositionb
P[ry+, HBLl](CyO-1)a 2215 0 200 30 (15%)P[ry+, HBLl](7M3-1) 1686
0 150 26 (17%)P[ry+, HBLl](CyO-2) 1536 0 50 9 (18%)
aThe first designation in parentheses indicates the chromosomal
linkage of the insertion and the second the line number for that
chromosome.bGemnlines with transposition were those that gave rise
to one or more exceptional G2 w+ males.1637
" 11,
I:IA N " , .z q
Ci -A >
- 1"
-, " -Z-:1.
-
B.R.Calvi and W.M.Gelbart
Production of transposase under control of the hsp7OpromoterOur
inability to detect hobo transcript raised the possibilitythat
transcriptional control might be the basis for hobogermline
specificity. The rarity of hobo RNA could be theresult of a low
level of transcription only in the few germlinecells of the fly. If
this model were correct, then replacementof the hobo promoter with
an inducible promoter should leadto somatic as well as germline
activity. The 5' end of hobowas replaced by the heat-inducible
hsp70 promoter and 5'untranslated leader (UT). This construct,
P[ry+, HSH2](heat shock hobo2 or HSH2), removes the first 147 bp
ofhobo, and fuses the 5' UT of hsp70 to a point 40 bpdownstream of
the putative hobo TATA box (Figure 1).To test the ability of the
HSH2 construct to confer activity
in the soma, animals containing HSH2 and two copies ofhawi were
raised at 25° C and heat shocked at different timesof development.
The resulting adults were frequently mosaicin eye pigmentation,
indicating that HSH2 is able to producenovel hobo activity in
developing eye tissue (Figure 2 andTable II). Later heat shocks
resulted in higher frequenciesof mosaicism, ostensibly because
there are more cells thatcould have a mobilization event. Heat
shock of a 4-6 dayold population of larvae at 37° C for 1 h
resulted in 100%of the adult flies containing numerous small clones
(typically
Fig. 2. hobo expression in somatic tissue. Mosaic white
expressionresulting from mobilization of the mini-white marked hobo
element,hawl, by the hsp7O-hobo fusion, HSH2. The mosaic fly on the
leftcontains both hawI and HSH2 and resulted from a 1 h, 37° C
heatshock of a population distributed over 0-2 days of development
(at25° C, after egg deposition). On the right is a nonmosaic
sibling toindicate the pigment levels imparted by the two donor
hawl insertionson the X chromosome in this line. The observation
that mosaic flieshave some sectors which are lighter than that due
to the donor sitessuggests hawl is undergoing excision as well as
transposition.
one to a few ommatidia). Heat shock of a 0-2 daypopulation of
animals resulted in mosaicism in 44-50% ofadult flies (Table II).
Results based upon 0-2 day heatshocks were chosen for comparison of
the three independentpositions of HSH2 because they resulted in
larger clones thatare more easily detected, and yielded a frequency
of clones
-
hobo germline specificity
A,rl:.o-.e
= 7-." -.o
probe
7.46
4.40)
I15112 -_. 2.37
OP 1.35
9
* (1.24
ilt)1 41
Fig. 3. Northern analysis of HSH2 RNA. RNA prepared from
larvaecontaining HSH2 that had been heat shocked for 1 h at 37° C
wasprobed with the XhoI fragment of hobo (essentially all of hobo
exceptfor 286 bp 5' and 106 bp 3'). A single, full-length HSH2
transcriptthat copurifies with poly(A)+ RNA migrates at -2.5 kb.
Analysis ofRNA from animals given a mild pre-heat shock first to
abrogate theeffect of heat shock on splicing also reveals one
full-length transcript(unpublished data). Hybridization to
ribosomal protein 49 transcript(RP49) is shown as a loading
standard (Al-Atia et al., 1985). 'A+',1 Ag poly(A)+ RNA; 'Total', 3
Ag total RNA; 'A-', 0.5 /tgpoly(A)- RNA; 'M', RNA mol. wt markers
(BRL). The size in kb ofthe mol. wt markers is indicated in the
margin.
possibly regulation. RNase protection using a probecorresponding
to the region 90-609 of hobo on HSH2 RNAdetected only one
full-length species of transcript (Figure 4).
Southern analysis of RT-PCR products from an 1.2 kbregion
spanning the ORFO/ORF 1 stop codon, however, didreveal shorter
bands that hybridize to hobo probes in somecases (Figure 5).
Although we did not observe the samebands using plasmid or in vitro
transcribed RNA controltemplates, there are reasons for believing
these shorter bandsare not relevant to hobo splicing and
regulation. First, eventhough shorter species should have amplified
moreefficiently, re-amplification of the primary PCRs using thesame
or nested primers yielded only the full-length product.Second, the
intensity of the shorter products did not correlatewith genetic
function. In some PCR amplifications,HSH2-containing animals that
had not been heat shockedyielded more of the shorter products than
heat shockedsiblings (unpublished data); nonetheless, the frequency
ofsomatic transposition with heat shock was much greater
thanwithout heat shock (as great as three orders of magnitudefor
heat shock during 4-6 days of development). Althoughanimals were
frozen immediately after heat shock for RNApreparation, the RNA
profile at later times was probably
Fig. 4. RNase protection analysis of the region spanning
theORFO/ORFI junction in HSH2 RNA. (A) Representation of the
probeused in this analysis. A 531 bp 32P-labeled antisense RNA
probe wastranscribed in vitro from an HFL1 subclone. The 531 bp
probecorresponds to the region between positions 609 and 90 of HFL1
andalso contains 11 bp of 5' polylinker sequence. Full-length
protection ofHSH2 RNA would give a 463 bp fragment because of
digestion ofpolylinker sequences and the region of the probe from
position 90 to147, the point of fusion of hobo to the hsp7O 5'
untranslated leader.Probe sequence corresponding to HSH2 is
represented as a horizontalline, and other sequences of the probe
as diagonal lines. (B) RNaseprotection reveals one unspliced
transcript species in the ORFO/ORFIregion. The size of the
protected fragment is that predicted by thereasoning in (A). Total
RNA was prepared from larvae which hadbeen heat shocked with a mild
pre-heat shock. Faint lower mol. wtbands are not an indication of
splicing because they are also present insamples derived from the
transformation host lacking HSH2(unpublished data). 'RNase-':
undigested full-length probe;'RNase+', RNase-digested probe
incubated with 1 itg HSH2 RNA;'M', mol. wt standards. Sizes of
standards in base pairs are indicatedin the margin.
similar because other transcripts unspliced during heat shockare
not spliced upon recovery (Yost and Lindquist, 1988).Thus, there is
no compelling evidence for splicing of ORFOto ORF1 as the mechanism
of hobo regulation.
1639
147
HSH227791
hsp7O
147 630
90
B +
710
489
404
_ 367
242
-
B.R.Calvi and W.M.Gelbart
A14-
II 142-N,
419S1Il-.4t 111 241) -
-1,i. 1111-,
.4'
.114-;
Fig. 5. RT-PCR analysis of HSH2 RN}'ORFO/ORFI junction. (A)
Schematic relanalyzed in the RT-PCRs. Below HSH2approximate
positions of primers used iImargins the sizes of the full-length
prodbeen treated with DNase was incubatedH1428- with or without
reverse transcrtranscript in a region that is highly
constransposase and therefore is likely part cPrimary
amplifications rely on a sense phsp7O-7+, and a hobo antisense
primer1.19 kb product if full-length. The samehsp7O-3+ and H1249-,
which would y1.15 kb, were used for reamplification cprimers are
named for the nucleotide pothe most 5' base of the primer is
identicRT-PCR products. Blots of PCRs wererepresenting virtually
the entirety of hobprimary amplifications using the
primers'Control': pHSH2 plasmid control templRNA derived from HSH2
lines given asevere heat shock either with (RT+) ortranscriptase
added before PCR. 'cn;ry':from a line lacking HSH2 with
(RT+);transcriptase. This is the line used as a IHSH2 and is devoid
of other hobo elemwere reamplified using either the same IH1271-
(lanes 6-8), or the nested prir(lanes 9-11). Secondary
amplificationswith reverse transcriptase did not yield 1data).
Although some reactions give smaproducts, they are not likely to be
derivspecies. For example, although primaryspecies for HSH2 RNA
(lane 2), only tiusing the same primers and conditions o10). The
position and size in kb of mol.blots are shown in the margins.
Use of a novel reporter to detect low level hobopromoter
activity
Ilxpcctid 1.11li-lll II ikb) To test further the hypothesis that
hobo germline specificityis due to a transcriptional mechanism, it
would be desirableto control the expression of a reporter gene with
the hobopromoter. Presumably because of the low level activity
of
1 l lp the hobo promoter, traditional reporter gene constructs
havefailed to yield detectable activity (unpublished data). We
thus
I.lI-10explored the use of a potentially more sensitive
reporter, Ptransposase.The activity of the natural P element in the
germline is
easily detected even though the transcript encoding functional1
ii :,< }, transposase is exceedingly rare (Laski et al., 1986;
Rio
et al., 1986). Additionally, although the P promoter isweakly
active in the soma and germline, removal of theregulated intron of
P in the construct A2-3 results in a highlevel of P transposase
activity in both tissues. Thus, becauseof the ease of detection at
low transposase concentrations,and the potential to monitor
activity in the soma and germlineby assaying P mobilization, we
chose A2-3 P transposaseas a reporter for the hobo promoter.We
replaced the P promoter in A2-3 with the putative hobo
promoter. Specifically the 1-147 fragment of hobo, frombp 1 of
hobo to 40 bp downstream of the hobo TATAconsensus, was fused to bp
69 of P, 4 bp downstream ofthe P TATA consensus (Kaufman and Rio,
1991). Becauseof the potential of this fusion to produce P
transposase,introduction into the genome via P
element-mediatedtransformation may have resulted in unstable
transformants.
i. We thus chose hobo-based transformation for introductionof
the A2-3 reporter into the genome. It was not known,however,
whether the 1-147 fragment contained sufficient
in the region spanning the 5' sequences required in cis for hobo
transformation. Wepresentation of the region therefore inserted the
hobo-P fusion into a complete hoboare shown the names and
transformation vector that contains a larger 5' end, and that
n this analysis and in the is marked with the rosy+ gene. This
construction, H[ry+,lucts expected. RNA that had HA2-3] (HA2-3),
contains all sequences of the hobo element.with the antisense
primer 'riptase. H1428- primes hobo HA2-3 has two hobo 5' ends i
the same orientation, an outer;erved with Ac and Tam3 end
representing the first 995 bp of hobo sequence and an)f the
functional message. inner, 1-147 putative promoter fragment of hobo
which is?rimer in the hsp7O 5' leader, fused to A2-3 (Figure 6).
Thus, through one transformationH1271 - which would yield a
experiment we could obtain transformants that contain a 5'primers
or the nested primers
'ield a full-length product of end representing only the 1-147
fragment, or, if 1-147)f primary PCR products. hobo is not used for
integration, insertions that contain the 1-995sition within HFL1 to
which end as well. Of three independent transformants, all-al. (B)
Southern analysis of contained both the inner 1-147 and the outer
1-995 5' endprobed with an XhoI fragment of hobo within HA2-3.'o.
'1° PCR' (lanes 1-5):s hsp7O-7+ and H1271 -. We measured P
transposase activity in the germline andlate; 'HSH2': RT-PCR of
total soma for lines containing HA2-3 driven by the hobomild heat
treatment plus a promoter and for a line containing a highly active
insert ofwithout (RT-) reverse A2-3 with the P promoter resident at
cytological positionRnd without (RT-) reverse 99B (Robertson et
al., 1988) (Table Ill). In the germline,transformation recipient
for the frequency of X to autosome transposition of a
mini-whiteents. '2° PCR': primary PCRs marked P element was reduced
several-fold for two linesprimers, hsp7O-7+ and and 20-fold for one
line of HA2-3 relative to A2-3(99B).Of samples not pre-incubated In
eye tissue, however, the activity of HA2-3 was reducedhybridizing
bands (unpublished at least 1000-fold relative to A2-3(99B). This
is aaller as well as full-length conservative estimate of the
difference because most of theed from spliced mRNA G1 flies in the
A2-3(99B) cross contained multiple mosaicamplification gives
shorter patches. Comparison of the ratio of P transposase activityr
nested primers (lanes 7 and in the soma versus the germline for
A2-3 with that of HA2-3.wt markers for the separate allowed
estimation of the relative contribution of the hobo
promoter to germline specificity (Table E). This comparison
1640
I!Iisi)7(),-9)I
-m-
-
hobo germline specificity
indicates replacement of the P promoter with hobo sequence A
test for Ha2-3 activity in tissues of the adultis sufficient to
confer a strong germline bias to A2-3 P cuticletransposase
activity. These results argue strongly that hobo We wished to ask
if the low expression observed for HA2-3activity is restricted to
the germline by a transcriptional in the eye would extend to other
somatically derived tissues.mechanism. Thus, HA2-3 activity in
tissues leading to adult cuticle was
tested by assaying P-mediated instability at the singed (sn)5\c
: locus. The mutant phenotype of snW is a slightly bent adult
9t\-]|{1{1 i;ll'i bristle. In the presence of P transposase,
which mobilizesP elements resident at the locus, the snw allele is
highlyunstable and can become more extreme (sne) resulting in
an
II+01x' 0 ~-'-'' E]+ extremely shortened bristle, or revert to
wild-type (sn+)(Engels, 1979, 1984; Robertson et al., 1988). We
chose toscore only sne bristles because the sn+ phenotype
lI1,i+. I1A 51l'. ' ' *1""! EN _ + overlapped that of snw in our
genetic background. When-:~v.rv_e>1: ^ snW' flies were crossed
to A2-3(99B), 60% of the bristlesscored were sne in the first
generation, indicative of P
Fig. 6. Schematic representation of hobo sequences fused to A2-3
P transposase activity in the soma (Table IV). In
contrasttransposase and their activity in the germline and soma.
Here all . . .
I
sequences of the vector integrated into the genome are shown
because HA2-3 is much less active than A2-3(99B) in its somaticof
their potential to affect expression. The putative hobo promoter
destabilization of snW. Among the three separate genomic(positions
1-147 of HFL1, 40 bp 3' of the TATA consensus) is fused positions
of HA2-3, only one bristle of 450 scored wasto position 69 of A2-3,
a P element derivative which lacks the possibly sne. In the
germline, however, the ability ofregulated intron, and therefore is
potentially active in germline and HA2-3 to induce sne flies was
only reduced to 40-73%soma. This fusion removes the TATA consensus
in the A2-3 element. hat fo (Table Ia) wecocldH[ry+, HA2-3l
contains this fusion in a rosy+ (y+) marked hobo that for A2-3(99B)
(Table IV). Thus, we conclude thatthetransformation vector. The
1-147 fragment of hobo is in the same somatic quiescence conferred
by hobo cis sequences is notorientation as the hobo vector. H[ry+,
HA2-3Inv] is identical to peculiar to eye tissue, but probably
represents a true somaH[ry+, HA2-3] except that the hobo
promoter-52-3 fusion is in germline dichotomy in hobo
expression.opposite orientation relative to the remainder of the
hobo vector. Because HA2-3 contained sequences representin both
theAlthough the relative activity in the soma is designated as a
'-', there Bis a low level of somatic activity detected (Tables III
and IV). ry+ 1- 147 and 1-995 ends of hobo, it was not known
ifsequences are shown foreshortened. hobo sequences are represented
as transcription began within the distal or proximal 5' end
ofshaded boxes. Above and below are the nucleotide coordinates for
the this construct. If transcription began in the distal 1
-995extent of hobo sequences in the elements. Arrowheads within
theboxes represent the terminal inverted repeats. The arrow below
the
sequence indicates the proposed transcriptional initiation site
within 5' leader which could potentially mediate a post-hobo.
transcriptional regulation. The low level of activity of the
Table Ill. P[w+] mobilization assay of HA2-3 in the soma and
germline
Transposase source Somatic activity (S) Germline activity (G)
G/S ratio
Adults Mosaic Gennlines Germlines withscored adults scored
transpositionb
P[ry+, A2-3](99B) 275 275 (100%) 20 15 (75%) 0.75H[ry+,
HA2-3](2-I)a 1975 1 70 8 (11%) 2.2 x 102H[ry+, HA2-3](2-2) 1696 2
70 10 (14%) 1.2 x 102H[iy+, HA2-3](3-1) 1390 0 70 3 (4%)
aThe first number in parentheses designates the chromosomal
linkage of the insertion and the second the line number for that
chromosome.bGermrlines with transposition were those that gave rise
to one or more exceptional G2 w+ males.
Table IV. snw assay of HA2-3 and HA2-3Inv in the soma and
germline
Transposase source Soma Germline
Bristles sne bristles Germlines Germlines withscored tested sne
progeny
P[ry+, A2-3](99B) 150 90 (60%) 16 16 (100%)H[ry+, HA2-3](2-I)a
150 1 13 7 (54%)H[ry+, HA2-3](2-2) 150 0 11 8 (73%)H[ry+,
HA2-3](3-1) 150 0 10 4 (40%)H[ry+, HA2-3Invl(3-1) 262 0 20 18
(90%)H[ry+, HA2-3Inv](2-1) 274 0 20 7 (35%)H[ry+, HA2-3Inv](2-2)
176 0 20 9 (45%)
aThe first number in parentheses designates the chromosomal
linkage of the insertion and the second the line number for that
chromosome.
1641
-
B.R.Calvi and W.M.Gelbart
hobo promoter did not allow us to locate the start
oftranscription by standard primer extension experiments. Wethus
constructed a second version of HA2-3, H[ry+,HA2-3Inv]
(HA2-3Inverted or HA2-3Inv) in which the1-147 fragment of hobo and
A2-3 P sequences wereinserted in the opposite orientation relative
to other hobosequences within the transformation vector (Figure 6).
Linescontaining HA2-3Inv displayed germline specific activity
atfrequencies comparable to HA2-3 (Table IV). Thus, theseresults
suggest that transcription is initiated within the 1-147fragment of
hobo containing the TATA and CAATconsensus, and that the HA2-3
transcript probably containsat most only a few bases of hobo
sequence.
DiscussionThe results presented here suggest that the primary
basis forthe germline restriction of hobo activity is the
limitation oftransposase to this tissue through transcriptional
control. Thisis evidenced by the ability of the hsp70 promoter to
directhobo activity in the soma and hobo sequences to
confergermline bias to the reporter A2-3. Although
formallypossible, we think it unlikely that the germline
specificityof HA2-3 is due to a post-transcriptional regulation
becauseits transcript probably contains only a few base pairs of
hobosequence. In addition, the absence of evidence for splicingof
the hobo transcript suggests that hobo germline specificityis not
due to a tissue specific mRNA species. Although wecannot eliminate
the possibility that there may be additionalpost-transcriptional
controls on hobo activity, the results withHA2-3 and HA2-3Inv
suggest that transcription is theprimary limitation to expression
in the soma. If there areadditional controls, they must be
different in some respectfrom those for the P element because
hsp70-hobo fusionsdisplay activity in the soma but hsp70-P fusions
do not.The I element, a retrotransposable element in the fly,
is
also known to have germline restricted activity (Buchetonet al.,
1984; Bucheton, 1990). The full-length, transposition-intermediate
RNA of this element is restricted to the femalegermline
(Chaboissier et al., 1990). This is probably dueto the restriction
of I element promoter activity to this tissue(Lachaume et al.,
1992; McLean et al., 1993). Thus thetranscriptional mechanism of
hobo may be similar to thatused by the I element. We find no
sequences that are similarbetween the promoters of the two
elements, however(unpublished data).
Sequences responsible for hobo expressionThe results with the
A2-3 reporter suggest that transcriptionbegins within the 1-147
fragment of hobo which containsthe TATA and CAAT consensus. Other
sequences of hobocontained within the transformation vector,
however, maycontribute to hobo transcriptional control. Given that
the1-147 fragment within HA2-3Inv is in opposite
orientationrelative to the transformation vector, if there are
othersequences within hobo that contribute, they must be able toact
somewhat independently of distance and orientation.Although larger
numbers are needed, the observation of
rare somatic events for HA2-3, but not P-HBL1, suggeststhat the
altered organization of hobo sequences within HA2-3may lead to a
detectable but low level of transcription inthe soma.
Alternatively, given steps subsequent totranscription are different
for the P-HBLl and HA2-3 activity
1642
assays, there may be additional, minor,
post-transcriptionalcontrols that act to reduce further the
activity of the intacthobo element.
The nature of hobo transcriptional controlIt has been shown that
transposition of P elements in thesoma is deleterious (Engels et
al., 1987; Woodruff, 1992).If there are selective pressures against
hobo somaticmobilization, the efficacy of transcriptional control
as ameans to restrict expression to the germline depends on
theefficiency of protection from activation by host enhancersand
promoters that act in somatic tissues. Protection againstfortuitous
activation due to genomic position is a recurringtheme in mobile
element regulation (Kleckner, 1990). Wethink it likely, therefore,
that hobo transcription is inhibitedby a negatively acting
mechanism in the soma. In fact,although we observe variable levels
of germline activity fordifferent insertions of both P-HBL1 and
HA2-3, we havenever observed a stimulation of activity in the soma
due togenomic position. Tests of other P-HBLl lines, and surveysof
wild-type strains containing multiple hobo elements fortheir
ability to mobilize hawi, have failed to reveal any withhobo
activity in the soma (unpublished data).Because the mini-white gene
within hawl is sensitive to
genomic position effects, the absence of hobo activity in
thesoma cannot be due to sequences analogous to those foundat the
hsp70 locus and in the gypsy mobile element that areable to act as
buffers to genomic position effects (Kellumand Schedl, 1992;
Roseman et al., 1993). Whatever themechanism of transcriptional
regulation, the observation thathobo activity in the male and
female germline can result inclusters of exceptional progeny
suggests that the hobopromoter is active early during premeiotic
germ cell divisions(Blackman et al., 1987; Ho et al., 1993).
Are natural hobo elements occasionally active in thesoma?There
have been several observations which suggest that insome strains
hobo elements may infrequently transpose inthe soma. In our early
work characterizing genetic instabilityat the dpp locus due to hobo
activity, in one line individualsmosaic for appendage defects
characteristic of dpp mutantswere observed at very low frequency
(Blackman and Gelbart,1989). This strain no longer displays this
property. We couldnot determine whether these were
transposase-dependentevents because the presence of numerous,
unmarkedelements in these strains did not allow for controls in
theabsence of transposase. Similarly, reports that in some
strainsdifferent nuclei in the same salivary gland contain
different,potentially hobo-mediated, chromosomal
rearrangementscannot be unambiguously ascribed to hobo
transposaseactivity (Lim, 1981; Yannopoulos et al., 1983).An
additional suggestion that hobo may be occasionally
active in somatic tissue comes from the observation
ofpolymorphic hobo positions in different salivary gland nucleiof
the same animal visualized by in situ hybridization topolytene
chromosomes (Kim and Belyaeva, 1991). This wasobserved in a
mutagenized strain (MS) which was isolatedon the basis of its high
spontaneous mutation rate [for reviewsee Ilyin et al. (1991)]. It
may be that this strain is mutantfor factor(s) that directly or
indirectly mediate hobotranscription in the soma. When the MS
strain is crossedto strains containing hawl, however, we do not
observe
-
hobo germline specificity
mosaic eye pigmentation. We find no evidence for adominant or
recessive mutation that permits hobo somaticmobilization
(unpublished data). It is possible that salivarygland cells are
permissive for hobo mobilization in thesestrains whereas eye tissue
is not. Alternatively, the differentstrains used to test hobo
activity in the in situ hybridizationand haw 1 assays may differ
with regard to additional factorsthat affect hobo somatic
activity.
While we have never observed high levels of expressionin the
soma due to genomic position, infrequently somepositions of hobo
that place it near particularly strongenhancers may result in
somatic transposase production. Dueto high instability in these
lines, mobilization of thetransposase-producing element would then
result in loss ofsomatic activity and selective advantage over
siblings thatretain mobilization in the soma. This would explain
the rare,transient observations of somatic hobo activity.
Additionally,some strains may contain mutations within host genes
orhobo elements that result in somatic activity. These mutationsmay
be similarly transient due to instability or selectivepressures
against high levels of somatic hobo activity.
The experiments described here provide a beginning
tounderstanding the mechanism by which hobo mobilizationis
restricted to the germline. Regulatory pathways leadingto hobo
germline specificity probably also control theexpression of host
genes. Continued investigation of thismechanism may aid in
elucidating the molecular distinctionsbetween soma and
germline.
Materials and methodsTransformation and genetic testsAll basic
strains are described in Lindsley and Zimm (1992). Two
strainsdevoid of hobo elements (E strains), cn; ry42 or y w67c23,
were used astransformation recipients for ry+ marked or w+ marked
elementsrespectively. P element based transformations were
essentially as describedby Spradling and Rubin (1982) except that
puchs7rA2-3 was co-injected asthe source of transposase (provided
by D.Rio). hobo element basedtransformations were as described by
Blackman et al. (1989), except thatthe source of hobo transposase,
HBLl, was used.hobo transposase tests involving mobilization of
hawl from the X
chromosome were essentially as described by Calvi et al. (1991).
For mostheat shocked versions of this cross, GI progeny distributed
over 2 daywindows of development were heat shocked in shell vials
at 37° C for 1h in a water bath (10 min to achieve temperature + 60
min heat shock),after which the vials were cooled to 25° C in a
water bath. All otherdevelopment was at 25° C. Both eyes of the
ensuing adult flies that containedtransposase and hawl were scored
for mosaic eye pigmentation at 16 xmagnification. The potential for
mosaicism segregated with the chromosomecontaining the hsp7O-hobo
fusions.Sources of P transposase were tested in two ways. The first
was based
on the mobilization of P[w+, newt] (kindly provided by
J.Sekelsky,unpublished data), a highly mobile P element marked with
the mini-whitegene that is resident on an X chromosome containing
the mutations y andw. In the Go, P transposase was introduced via
the male and P[w+. newt]via the female. Both eyes of the GI were
inspected for mosaic eyepigmentation. Single male flies containing
the source of transposase andP[w+, newt] element were crossed to
two w sn3 tester females in vials for6 days at 25° C. The vials
were then scored on days 16-17 for theappearance of transpositions
(sn w+ sons) among the G2.The second test was based on singed-weak
(snW) mutability (Engels,
1984; Robertson et al., 1988). P transposase-containing males
were crossedto y snw; bw; st females and the male GI progeny
hemizygous for the snlocus were scored for bristle defects. The
left and right posterior scutellarbristles were scored for each fly
examined. Only sne defects were recordedbecause sn+ and snw
phenotypes closely resembled each other in ourgenetic background.
For sn genmline tests, males containing the source ofP transposase
and y sn* chromosome were crossed singly in vials to twoattached-X
females [C(J)DX, y wfl. GI adults were cleared from the vials
on day 6 and patroclinous male G2 progeny were scored for
nonmosaicsne bristle phenotype on days 16-17. All crosses were
reared at 25° C.
RNA analysisRNA was extracted from various stages using a hot
phenol method asdescribed by Brown and Kafatos (1988). Poly(A)+ RNA
was purified byone passage over a poly(T) - Sepharose column
according to the conditionsof the supplier (BMB). For HSH2
heat-shocked RNA, animals were heatshocked for 1 h at 37° C and
then immediately frozen in liquid N2. Whereindicated animals were
given a mild heat shock of 35° C for 30 min, followedby 3 h
recovery at 25° C, and then severe heat shock of 37° C for 1 h.
Thispre-heat shock has been shown to reduce the inhibition to
splicing observedfor severe heat shocks alone [for review see Yost
et al. (1990)].For Northerns, gel electrophoresis in
formaldehyde-containing gels and
transfer to nylon membranes were as described by Lehrach et al.
(1977)and Rabinow and Birchler (1989). pRG2.6X, an XhoI subclone
whichrepresents virtually the entirety of hobo (Blackman et al.,
1987), was usedfor the generation of 32P-labeled hobo probes by
random oligopolymerization using random hexamers (Feinberg and
Vogelstein, 1983,1984). Hybridization and washes were performed at
65° C in Church Bufferas described (Church and Gilbert, 1984).
Northern blots were reprobed withrandom oligo generated probes
prepared from the plasmid rp49 to detectribosomal protein 49 (RP49)
message to control for loading (Al-Atia, 1985).RNA markers (BRL)
were used as size standards.RNase protection and probe preparation
were essentially as described
(Krieg and Melton, 1987) except for the following modifications.
The plasmidHA609 was used as template for in vitro transcription of
32P-labeledantisense probe. HA609 was created from HFL1 by
digestion with PstIand SacI, making the ends blunt with T4
polymerase, and gel purificationand self-ligation of the vector/5'
hobo fragment. Thus HA609 is deletedfor all sequences 3' of
position 609 of HFLI until the Sacl site within thepolylinker.
HA609 template was linearized by digestion with RsaI at position90
of HFL1 sequence. Probe was synthesized by T7 RNA
polymeraseresulting in a 531 bp antisense transcript which is
identical to HFL1 fromposition 90 to 609 and includes 11 bp of
polylinker sequence. Full-lengthin vitro transcribed RNA probe was
purified on 4% acrylamide-6 M ureagels and eluted directly into
formamide hybridization buffer. Markers wereprepared by digestion
of pBS-KS (Stratagene) with Hpall and labeled byfilling the ends
using the Klenow fragment of DNA polymerase I and allfour [32P]NTPs
at 800 Ci/mmol.RT-PCR of HSH2 RNA was essentially as described by
Fuqua et al.
(1990), with the following modifications. 20 ytg total RNA was
DNase-treated at 37° C for 30 min in a 100 1l reaction containing
30 U of RNase-free DNase I (BMB), 80 U of RNAsin (Promega), 10 mM
DTT, 0.1 Msodium acetate pH 5, 5 mmol MgSO4, after which RNA was
phenolextracted twice and ethanol precipitated. 1 Ag of
DNase-treated RNA wasdenatured for 3 min at 100° C and incubated
with or without 60 U of MMLVreverse transcriptase in a 10 A1
reaction containing 50 mM KCI, 3 mMMgCl, 10mM Tris-HCI pH 9,0.1%
Triton X-100, 0.75 mM each dNTP,I U/1l RNAsin, 10 mM DTT, 1.25
pmol/Al primer, for 1 h at 370Cfollowed by 30 min at 45° C. The 10
1l reverse transcription was heatedto 100° C for 3 min and added
direcdy into 40 $1 of prewarmed PCR solution.The final
concentrations were 50 mM KCI, 2 mM MgCl2, 10 mMTris-HCI pH 9, 0.1%
Triton X-100, 150 mM each NTP, sense andantisense primer at 0.25
-0.5 pmol/4l each, and 2.5 U Taq polymerase.PCR typically involved
25-30 cycles each consisting of 1 min at 940C,
min at 50° C, and 2 min at 72° C. The hobo primer, H1271-,
andhsp7O-7 +, a primer corresponding to sequence within the hsp 70
untranslatedleader present in HSH2, were used for primary
amplifications. For secondaryamplifications 1 1d of the primary PCR
products were reamplified in 501I reactions using either the same
primers, H1271 - and hsp7o-7 +, or thenested primers, H1249- and
hsp7o-3 +, corresponding to hobo and hsp 70sequence respectively.
hobo primers were named for the position withinHFL1 for which the
most 5' nucleotide of the primer corresponds, with+ and - symbols
to designate sense and antisense primers respectively.The sequences
of the primers are as follows: hsp7o-3 +,
GGCGGTACC-GAATACAAGAAGAGAAACTCTG; hsp7o-7+,
GCCAAGAAGTAA-TTATTGAATAC; H 1428-, GTAGTTGGAGTTCCATCTAGTCGG;H1271-,
GGTTGGATCCTGACAACAGGTGACTGCTAC;
H1249-,TAAACGGGTATTGCCCTCTAAAGCC.PCR products derived from hobo
were identified by probing Southern
blots of the PCRs with 32P-labeled pRG2.6X. Hybridization of
bands tohobo probes in lanes representing reactions with reverse
transcriptase, butnot in those without reverse transcriptase,
confirms they are derived fromhobo RNA. Full-length PCR products
were subcloned and restrictionmapped, confrming they do not contain
deletions. Multiple attempts to clone
1643
-
B.R.Calvi and W.M.Gelbart
the shorter primary PCR products that do not reamplify have
beenunsuccessful.
DNA cloningProcedures for plasmid constructions were essentially
as described (Sambrooket al., 1989). HSH2 was constructed by
inserting the FspI fragment of HFLI(from 147 to 2791) into a ry+ P
element vector containing the hsp7Opromoter which was prepared by
digestion with Notl and made blunt-endedby the Klenow fragment of
DNA polymerase I. This ry+ P element vectorcontaining the hsp7O
promoter, P[ry+, hs], had been created by insertingthe Sail-NotI
fragment containing the hsp7O promoter from pCasper-hs(provided by
C.Thummel) into the NotI and Sail sites of pDM30 (Mismerand Rubin,
1987).HA2-3 was constructed by first ligating the Kpnl-FspI
fragment of HFL1
and a BanHI-ApaLI filled in fragment from puchsirA2-3 (provided
byD.Rio) into the KpnI- and BamHI-cut vector pmartini. pmartini was
createdby inserting the polylinker from pPoly IE-I (Lathe et al.,
1987) into thevector pBS-KSII+ (Stratagene). A NotI fragment
containing the hobo-Pelement fusion was excised from this vector
and ligated into the NotI siteof the ry+ marked hobo vector H[ry+,
HFLN]. H[ry+, HFLN] is aderivative of HFL1 in which the Sall site
has been changed into a NotIsite by the addition of a linker, and
which has the 7.2 kb HindIIl ry+fragment from Car20 (Rubin and
Spradling, 1983) ligated into the uniqueHindIII site of hobo. In
HA2-3Inv the NotI fragment containing positions1 -147 of hobo fused
to A2-3 is in inverted orientation relative to the
hobotransformation vector.
AcknowledgementsThe authors wish to thank the following
individuals. D.Rio for helpful adviceand puchs7rA2-3; C.Thummel for
pCasper-hs; S.Bray, T.Hsu andG.Tzertzinis for advice on PCR;
K.Mowry and M.Mortin for advice onRNase protection; J.Sekelsky for
the gift of P[w+, newt], discussions andcomputer maintenance;
D.Smith for early attempts with reporters;M.Sanicola and S.Findley
for vector construction; D.Coen for helpful advice;F.Kafatos,
W.Bender, C.Swimmer and R.Padgett for a critical reading ofthe
manuscript; E.Chartoff and Y.Kraytsberg for technical
assistance;L.Lukas, D.Lukas and J.Gargano for fly support. B.R.C.
was supportedby a NSF predoctoral fellowship and a PHS predoctoral
traineeship ingenetics. This research was supported by a PHS grant
to W.M.G.
ReferencesAl-Atia,G.R., Fruscoloni,P. and Jacobs-Lorena,M.
(1985) Biochemistry,
24, 5798-5803.Blackman,R.K. and Gelbart,W.M. (1989) In Berg,D.E.
and Howe,M.M.
(eds), Mobile DNA. American Society for Microbiology,
Washington,DC, pp. 523-529.
Blackman,R.K., Grimaila,R., Koehler,M.M.D. and Gelbart,W.M.
(1987)Cell, 49, 497-505.
Blackman,R.K., Koehler,M.M.D., Grimaila,R. and Gelbart,W.M.
(1989)EMBO J., 8, 211-217.
Brown,N.H. and Kafatos,F.C. (1988) J. Mol. Biol., 203,
425-437.Bucheton,A. (1990) Trends Genet., 6, 16-21.Bucheton,A.,
Paro,R., Sang,H.M., Pelisson,A. and Finnegan,D.J. (1984)
Cell, 38, 153-163.Calvi,B.R., Hong,T.J., Findley,S.D. and
Gelbart,W.M. (1991) Cell, 66,465-471.
Chaboissier,M.-C., Busseau,I., Prosser,J., Finnegan,D.J. and
Bucheton,A.(1990) EMBO J., 9, 3557-3563.
Church,G.M. and Gilbert,W. (1984) Proc. Natl Acad. Sci. USA,
81,1991-1995.
Engels,W.R. (1979) Proc. Natl Acad. Sci. USA, 76,
4011-4015.Engels,W.R. (1984) Science, 226, 1194-1196.Engels,W.R.
(1989) In Berg,D.E. and Howe,M.M. (eds), Mobile DNA.
American Society for Microbiology, Washington, DC, pp.
437-484.Engels,W.R., Benz,W.K., Preston,C.R., Graham,P.L.,
Phillis,R.W. and
Robertson,H.M. (1987) Genetics, 117, 745-757.Feinberg,A.P. and
Vogelstein,B. (1983) Anal. Biochem., 132, 6.Feinberg,A.P. and
Vogelstein,B. (1984) Anal. Biochem., 137, 266.Fuqua,S.A.W.,
Fitzgerald,S.D. and McGuire,W.L. (1990) BioTechniques,
9, 206-211.Ho,Y.T., Weber,S.M. and Lim,J.K. (1993) Genetics,
134, 895-908.Ilyin,Y.V., Lyubomirskaya,N.V. and Kim,A.I. (1991)
Genetica, 85,13-22.
Kaufmnan,P.D. and Rio,D.C. (1991) Proc. Natl Acad. Sci. USA,
88,2613-2617.
Kaufnan,P.D. and Rio,D.C. (1992) Cell, 69, 27-39.Kellum,R. and
Schedl,P. (1992) Mol. Cell. Biol., 12, 2424-2431.Kim,A.I. and
Belyaeva,E.S. (1991) Mol. Gen. Genet., 229, 437-444.Kleckner,N.
(1990) Rev. Cell Biol., 6, 297-327.Krieg,P.A. and Melton,D.A.
(1987) Methods Enzymol., 155, 397-415.Lachaume,P., Bouhidel,K.,
Mesure,M. and Pinon,H. (1992) Development,
115, 729-735.Laski,F.A. and Rubin,G.M. (1989) Genes Dev., 3,
720-728.Laski,F.A., Rio,D.C. and Rubin,G.M. (1986) Cell, 44,
7-19.Lathe,R., Vilotte,J.L. and Clark,A.J. (1987) Gene, 57,
193-201.Lehrach,H., Diamond,D., Wozney,J.M. and Boedtker,H.
(1977)
Biochemistry, 16, 4743.Lim,J.K. (1981) Cold Spring Harbor Symp.
Quant. Biol., 45, 553-560.Lim,J.K. (1988) Genetics, 85,
9153-9157.Lindsley,D.L. and Zimm,G.G. (1992) The Genome of
Drosophila
melanogaster. Academic Press, San Diego, CA.McClintock,B. (1948)
Carnegie Inst. Wash. Year Book, 47, 155-169.McClintock,B. (1952)
Carnegie Inst. Wash. Year Book, 51, 212-219.McLean,C., Bucheton,A.
and Finnegan,D.J. (1993) Mol. Cell. Biol., 13,
1042-1050.Mismer,D. and Rubin,G.M. (1987) Genetics, 116,
565-578.O'Hare,K. and Rubin,G.M. (1983) Cell, 34,
25-35.Pirrotta,V., Steller,H. and Bozzetti,M.P. (1985) EMBO J., 4,
3501-3508.Rabinow,L and Birchler,J.A. (1989) EMBO J., 8,
879-889.Rio,D.C. (1991) Trends Genet., 7, 282-287.Rio,D.C.,
Laski,F.A. and Rubin,G.M. (1986) Cell, 44, 21-32.Robertson,H.M.,
Preston,C.R., Phillis,R.W., Johnson-Schlitz,D.M.,
Benz,W.K. and Engels,W.R. (1988) Genetics, 118,
461-470.Roseman,R.R., Pirrotta,V. and Geyer,P.K. (1993) EMBO J.,
12, 435-442.Rubin,G.M. and Spradling,A C. (1983) NucleicAcids Res.,
11, 6341 -6351.Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989)
Molecular Cloning. A
Laboratory Manual. 2nd edn. Cold Spring Harbor Laboratory Press,
ColdSpring Harbor, NY.
Siebel,C.W. and Rio,D.C. (1990) Science, 248,
1200-1208.Siebel,C.W., Fresco,L.D. and Rio,D.C. (1992) Genes Dev.,
6, 1386-1401.Smith,P.A. and Corces,V.G. (1991) Adv. Genet., 29,
229-300.Spradling,A.C. and Rubin,G.M. (1982) Science, 218,
341-347.Streck,R.D., MacGaffey,J.E. and Beckendorf,S.K. (1986) EMBO
J., 5,
3615-3623.Woodruff,R.C. (1992) Genetica, 86,
143-154.Yannopoulos,G., Stamatis,N., Zacharopoulou,A. and
Pelecanos,M. (1983)
Mutat. Res., 108, 185-202.Yannopoulos,G., Stamatis,N.,
Monastirioti,M., Hatzopoulos,P. and
Louis,C. (1987) Cell, 49, 487-495.Yost,H.J. and Lindquist,S.
(1986) Cell, 45, 185-193.Yost,H.J. and Lindquist,S. (1988) Science,
242, 1544-1548.Yost,H.J., Petersen,R.B. and Lindquist,S. (1990)
Trends Genet., 6, 223.
Received on November 23, 1993; revised on January 10, 1994
1644