Regulation of ribosomal RNA cistron number in a strain of Neurospora crassa with a duplication of the nucleolus organizer region

162

Btochimica et Btophysica Acta, 697 (1982) 162 169 F_lscvier Biomedical Pres~

BBA91072

REGULATION OF RIBOSOMAL RNA CISTRON NUMBER IN A STRAIN OF N E U R O S P O R A C R A S S A WITH A DUPLICATION OF THE NUCLEOLUS ORGANIZER REGION KARIN D. RODLAND and PETER J. RUSSELL

Biology Department, Reed College, Portland, OR 97202 (U.S.A.) (Received October 16th, 1981) (Revised manuscript received February 16th, 1982)

Key words: rRNA," Cistron; Nucleolus organizer region," (N. crassa)

Some progeny from a cross of the transiocation mutant T(VL--, IVL)AR33 with wild-type Neurosporacrassa are double nucleolus organizer (DNO) strains, usually displaying two distinct nucleolus organizer regions. The DNO strain is sterile but displays the same growth response as normal laboratory strains of Neurospora. We used DNA-DNA hybridization techniques to quantify the number of rRNA cistrons in the DNO mutant and its vegetative progeny. Comparisons of the rate of hybridization of genomic DNA from the parental AR33 strain and from the DNO strain showed that hybridization was more rapid for the DNO strain than for the parental strain. Successive vegetative progeny of the DNO strain displayed hybridization rates intermediate to those of the original DNO strain and the parental single nucleolus strain, indicating that the number of rRNA cistrons had decreased during vegetative propagation. Estimates of rRNA cistron number obtained from comparisons of the amount of single copy DNA and rDNA hybridized to genomic DNO and AR33 DNA at saturation indicate that the parental AR33 strain contains 225 copies of the rRNA repeat unit, while the DNO strain has approx. 440 copies. The number of rRNA cistrons decreases gradually in the successive vegetative progeny, approximating the parental haploid value by the eleventh vegetative transfer.

Introduction

The nucleolus and nucleolus organizer are closely associated with ribosome and ribosomal R N A biosynthesis in eukaryotes; the nucleolus organizer is the chromosomal region coding for rRNA while the nucleolus, which forms around the NO region, functions as the site of ribosome biosynthesis [1-4]. Nucleoli are cytologically prominent and closely associated with the nucleolus organizer [5,6], and these characteristics make the nucleolus and nucleolus organizer especially useful in molecular genetics. We have used both these attributes in a study of rRNA cistrons in strains of Neurospora crassa possessing visible alAbbreviation: DNO, double nucleolus organizer. 0167-4781/82/0000-0000/$02.75 © 1982 Elsevier Biomedical Press

terations in the number a n d / o r size of nucleoli present. Haploid throughout most of its life cycle, wild-.type strains of N. crassa have one nucleolus. Cytological studies have localized the nucleolus organizer region to the left arm of chromasome 2, which is linkage group V [6]. Gene cloning and restriction endonuclease mapping data indicate that the cistrons for 17 S, 5.8 S and 25 S rRNA are arranged in that order with interspersed spacer regions to form an 'rRNA repeat unit' of 8.1 kilobase pairs (Refs. 7, 8; R.L. Metzenberg, personal communication). DNA-DNA reassociation experiments indicate that there are 185 tandemlyarranged copies of the rDNA repeat unit in the Oak Ridge-derived wild-type strain of N. crassa, 74-OR23-1a [9,10].

163 Recently, Perkins et al. [11] described a translocation mutant strain of N. crassa, T(VL IVL)AR33, in which a segment of the left arm of linkage group V (containing the nucleolus organizer region) is translocated to the end of the left arm of linkage group IV. When this translocation strain is crossed with a normal strain, some progeny have either two nucleoli or a single enlarged nucleolus. These progeny are interpreted to be duplication strains which have two nucleolus organizer regions. We call these duplications double nucleolus organizer, or DNO strains. In the experiments described in this paper we have used DNA-DNA hybridization techniques to quantify the number of rRNA cistrons present in the DNO strain. As expected, we found that the DNO strains have twice the number of cistrons as the parental translocation strain. In addition, we investigated whether or not Neurospora has a regulatory mechanism for adjusting the rRNA cistron number to the 'normal' level upon successive vegetative transfers. The ability to adjust rRNA cistron has been demonstrated in other organisms, such as the bobbed ( bb ) mutants of Drosophila melanogaster which are initially deficient in rDNA compared to wild-type Drosophila, but which eventually display wild-type rDNA levels after several generations [12]. Our results indicate that Neurospora is able to regulate the number of rRNA cistrons present. Materials and Methods

Strains. The translocation strain AR33, carrying a caffeine-resistance marker and symbolized T(VL ---, IVL)AR33 caf-lRA (Fungal Genetics Stock Center No. 2021) was obtained from D.D. Perkins, Stanford University. A lys-1; cys-lO a (linkage group VL and IVL, respectively) strain was obtained from the same source. The normal sequence wild-type strains used for comparison were the Oak Ridge-derived strain 74-OR23-1a (obtained from R. Metzenberg, University of Wisconsin, Madison) and the St. Lawrence strain 74 A, further inbred in the laboratory of A.M. Srb (Cornell University) from whom it was obtained. Isolation numbers of the markers used are: caf-1 R, KH101; cys-lO, 39816; lys-1, 33933. Materials. Nitrobenzyloxymethylcellulose paper ('Transa-bind') was obtained from Schleicher and

Schuell, Keene, NH. Bovine pancreas deoxyribonuclease (DNAase I) was obtained from Worthington Biochemical Corporation, Freehold, N.J. Ribonuclease A (RNAase) from bovine pancreas and Protease V from Streptomyces griseus were obtained from Sigma Chemical Company, St. Louis, MO. DNA polymerase I from Escherichia coli was obtained from Bethesda Research Laboratories, Gaithersburg, MD. 32p-labeled dCTP was obtained from New England Nuclear, Boston, MA. The remaining chemicals were obtained from Sigma Chemical Company, St. Louis, MO. DNA isolation and purification. DNA was isolated from mycelia obtained from fresh conidial inoculations of Vogel's minimal medium [15] supplemented with 0.5 mg lysine/ml and incubated for 14h at 25°C. The mycelia were harvested by filtration and suspended in 3 vol. of a solution containing 1 M sorbitol/0.25M Li3EDTA, pH 7.5/20% (v/v) glycerol/0.5% (v/v) Triton X-100. The cells were lysed by freezing the suspension at - 7 0 ° C for 1 h and extruding the paste through a French pressure cell. Protease V and sodium dodecyl sulfate (SDS) were added to a final concentration of 250 /~g/ml and 1% (w/v), respectively, and the mixture was thawed at 48°C for 2 h. The macromolecules were precipitated with 2.5 vol. of ethanol at -20°C. A pellet was obtained by centrifuging for 10 min at 16000×g in a Sorvall RC-5 centrifuge, GSA rotor. The nucleic acids and proteins were dissolved in 10 ml high salt buffer (25 mM Tris-HC1, pH 7.4/100 mM NaC1/2 mM EDTA) and the remaining cellular debris was discarded. Nucleic acids were separated from proteins and other macromolecules using saturated ethanolic perchlorate (Wilcockson's reagent; Ref. 14). DNA was purified from RNA by incubating the nucleic acids in high salt buffer containing 250 /zg/ml pre-digested RNAase for 2 h at 37°C. The RNAase was removed by treatment with Wilcockson's reagent, and the DNA was collected by precipitation in ethanol at - 20°C, and centrifugation at 16000 X g for 10 min. The DNA was dissolved in low salt buffer (1 mM Tris-HC1, pH 7.4/1 mM NaC1/0.1 mM EDTA) and analyzed by electrophoresis in a 1% agarose gel. Comparison with a marker mix of HindlII and EcoRI digested ~ DNA (obtained from F. Blattner, University of Wisconsin, Madison) indi-

164 cated that DNA obtained using this procedure was approx. 12-14 kilobase pairs long and free of RNA contamination. Shearing the DNA by passing the solution through a 23-gauge needle five times produced fragments which were approx. 0.5-1 kilobase pairs long, and this treatment was used for preparation of the hybridization probes. The concentration of DNA in solution was determined by its absorption at 260 nm and confirmed by diphenylamine assay. DNA-DNA hybridization. Hybridization of genomic and cloned ribosomal DNA was performed by a modification of the 'Northern blot' procedure described by Alwine et al. [15] and Wahl [16]. DNA was covalently bound to diazobenzyloxymethyl cellulose paper which had been freshly prepared from the nitrobenzyloxymethyl form according to the protocol of Alwine et al. [15]. The diazobenzyloxymethyl paper was incubated for 24-36 h at 4°C in a solution containing 2 mg/ml DNA in a 1:4 mix of 50 mM sodium borate buffer, pH 8.0/dimethyl sulfoxide. The paper was then washed three times in distilled water, four times in 0.4 M NaOH (10 min at 37°C each wash) and three times in distilled water. Analysis of the DNA solution after the binding incubation indicated that 6.4 /xg DNA was bound/cm 2. The paper was stored at 4°C in 50% (v/v) form a m i d e / 0 . 7 5 M N a C I / 7 5 mM sodium citrate/0.2% (w/v) SDS. Hybridization was carried out in the same solution containing 10-300 ~ g / m l labelled probe DNA, and the paper and buffer were incubated at 37°C for 36 h to 5 days, depending on the concentration of DNA. Determination of the time course of hybridization for single copy DNA (plasmid pLHI 1, described in the plasmid section) indicated that hybridization was saturated at Cot values greater than 70 (9.9-+0.9 ng bound at 70 Cot vs. 10.0-+0.7 ng bound at 100 Cot, p>0.3). The hybridization experiments in which cistron number was determined were conducted at a Cot value of 90. The filters were counted in a Beckman Model 7500 LSC, and the quantity of DNA bound was calculated from the observed cpm/filter and the specific activity of the initial hybridization solution, determined under identical conditions. Plasmids. Three plasmids isolated from recombinant DNA clones were used as hybridization

probes: pMF2, a PstI fragment of N. crassa 74OR23-1a DNA cloned into pBR322 and which has been shown to contain the coding sequences for 17 S, 5.8 S and 25 S rRNA as well as some flanking sequences [7]; pVK88, a PstI-generated fragment cloned into the PstI site of pBR322; and pLHl 1, a BglI-BamHI fragment cloned into pBR322. pVK88 contains the single copy catabolic dehydrogenase gene qa-2 and the dehydroshikimic acid dehydrase gene qa-4 while p L H l l contains only qa-2 (Ref. 17; Giles, personal communication). pMF2 and pVK88 were obtained from R. Metzenberg, University of Wisconsin, Madison, and pLH11 was obtained from N. Giles, University of Georgia. Radioactive labelling of DNA. DNA was labelled in vitro by nick translation based on the procedure of Rigby et al. [18]. A 100 /xl reaction mixture containing 1-2 ~g DNA/100-150/xCi 32p-dCTP (high spec. act. in 10 mM Tricine)/50 mM TrisHCI, pH 7.1/10 mM Mg SO4/50 /~tg/ml bovine serum albumin/0.06% mercaptoethanol/10 mM dithiothreitol/15 /~M dATP/15 /.tM dTTP/ 15 /~M dGTP/2 units DNAase I/1.5 units E. coli DNA polymerase I were incubated at 14°C for 75 min. The reaction was stopped by adding 100/~1 50% sucrose/l% bromophenol blue/10 mM TrisHC1/10 mM NaCI/2 mM EDTA, pH 7.7. The labelled DNA was separated from the unincorporated nucleotides using a Sephadex G100 column. Specific activities of 1-3 - 1 0 7 cpm/mg DNA were obtained. Results

Production of the double nucleolar organizer duplication strain The generation of the duplication strain containing two nucleolus organizer regions (the DNO strain) is illustrated in Fig. 1 (see Perkins et al. [11], for a more complete discussion). A normal strain, in this case lys-1; cys-lO a, is crossed with the rearrangement strain T(VL---, IVL)AR33 caf1R A. One-quarter of the meiotic products contain lethal deficiencies. Of the viable spores produced, one-third are stable duplications; these are all lysine-requiring and cysteine-independent. The duplications are also heterozygous for the dominant caffeine-resistance marker. Perkins et al. [11] have

165

N lys-I;cys'lOo x

T(~]L"]~L)AR55cof'I R A

014-

,l

0.13-

+ I~_ cys-IO OI

£R

I

L_l---

0 - ....

,o,, 1 [c-*'-- ....

Meiotic

c°f-I$1

lVR

•

o

o

Q12-

°f'lR o

Ixliring

~R

~7 O, YL

1

IVR

I l Some progeny have two NO's

0.11-

....

°°"'I I°°'''" ~NO

sot

~L Fig. I. Meiotic (pachytene) pairing in a cross of T ( V L ~ IVL)AR33 with a normal (wild-type) strain. The viable meiotic products contain two nucleolus organizers per nucleus when the two lower chromosomes migrate to the same pole in anaphase I. Symbols: - - , linkage group V; . . . . . . , linkage group IV; ca]', caffeine resistant; cys, cysteine; lys, lysine; NO, nucleolus organizer; sat, satellite.

shown that such strains have two nucleoli. As with other duplications in Neurospora these duplication strains are largely infertile. Growth phenotype of DNO and wild-type N. crassa In order to determine whether the DNO strain is physiologically disadvantaged by the duplication, the initial growth response of the DNO strain was compared with that of a normal strain, 74 A. Fig. 2 shows that there is no significant difference in the growth rate of the D N O strain and 74A during a 6 h incubation at 25°C. This suggests that initial germination and growth of the D N O strain is not adversely affected by the duplication, and that its overall physiology remains within normal limits. This facilitates direct comparisons of DNA obtained from D N O and wild-type strains grown under similar conditions. Relative number of rRNA repeat units in DNO and normal strains: kinetics of D N A / D N A hybridization If the D N O mutant has an increased number of

0 . 1 0 "~ 0

f

J

J Tiffs[h)

Fig. 2. Growth response of wild type and DNO strains. Conidia

from the wild-type strain 74A-9 (Q) and the DNO strain (©) were incubated at 25°C for 6 h in Vogel's complete medium supplemented with 0.5 mg lysine/ml. Growth was monitored by measuringincreasesin turbidity (absorbanceat 450 nm) as a function of time.

rRNA repeat units associated with the presence of two nucleolus organizers, then hybridization reactions between genomic D N A and either rRNA or rDNA should proceed more rapidly for D N O genomic DNA than for normal genomic DNA. Furthermore, if subsequent vegetative transfers of the duplication strain involve a loss of rRNA repeat units, then the hybridization rate of the successive vegetative transfers should decrease, eventually approximating that of the parental translocation strain T(VL--, IVL)AR33. These changes in the D N A / D N A hybridization parameters should be reflected in both the quantity of r D N A hybridized to genomic D N A at saturation and in the apparent reassociation constant, k, observed for hybridization. Both parameters have been used by other investigators to estimate the number of rRNA cistrons present (Phillips et al., in maize [19,20]; Krumlauf and Marzluf in N. crassa [9,10]). We have used variations of both methods to quantify the amount of rDNA present. In the first set of experiments, varying concentrations of genomic DNA from a D N O strain and the parental translocation strain, AR33, were hybridized to cloned Neurospora rDNA (plasmid pMF2 obtained from R. Metzenberg) covalently

166

1816.

1.4-

~"~ 12 Z C3

~

J

08

o

>, 06 I

Od 02 ¸ o

o6~ o~o o4~ o~o o5~ o~o DNA

Col

Fig. 3. Double-reciprocal plot of hybridization rate vs. DNA

Cot. Varying concentrations of in vitro 32p-labelled genomic DNA from the parental T(VL~IVL)AR33 strain (O), the freshly generated DNO strain (O), and the fifth vegetative transfer of the DNO strain ( X ) were hybridized to cloned Neurospora rDNA (pMF2 obtained from R. Metzenberg) bound covalently to DBM-cellulose paper. Hybridization occurred in a solution containing 50% formamide/0.1% (w/v) SDS/0.75 M sodium chloride/75 mM sodium citrate, pH 6.5 during incubation at 37°C for 4 - 5 . l0 s s. Data from the linear portion of the hybridization curve were used to construct the double-reciprocal plot of hybridization rate vs. DNA Cot.

bound to diazobenzyloxymethyl paper. Hybridization was carried out at 37°C for 4 - 5 days until hybridization appeared to be saturated at the highest concentrations. Data from the linear portion of the hybridization curve were used to construct a double-reciprocal plot of D N A Cot vs. hybridization, and the horizontal intercept was used as an estimate of k. Fig. 3 shows the combined results of three such hybridization experiments using D N A from D N O progeny from the first vegetative transfer (T No. 1) and from the fifth vegetative transfer (T No. 5). D N A from the original parent T ( V L - > |VL)AR33 was used as a control. The data show that hybridization of genomic D N A to cloned r D N A is much more rapid

for the T No. 1 D N O strain than for the parental haploid strain, while the T No. 5 hybridizes at an intermediate rate. Estimates of k obtained from the horizontal intercept are approx. 8.7 M ~. s for the D N O strain and 3.95 M l . s ~ for the AR33 strain. The ratio of these k values is 2.2, indicating that the D N O strain has approx, twice as many cistrons as the parent strain. The intermediate values obtained for hybridizations involving T No. 5 show that many of the excess rRNA repeat units have been lost in five vegetative transfers, although the parental r D N A level has not yet been attained. While the preceding experiments indicate the relative amount of r D N A present in the D N O and parental AR33 strains, they do not allow precise determination of the number of r R N A cistrons present in the two strains. An accurate determination of the number of r R N A repeat units present in the D N O strain and its vegetative progeny would provide more information about the precision with which r R N A cistron number is regulated in Neurospora. Consequently, experiments were done to estimate r R N A cistron number. In these experiments we compared the total amount of r D N A hybridized to genomic D N O and AR33 D N A with the amount of single copy D N A hybridized under identical conditions. The genomic D N A was bound to diazobenzyloxymethyl paper and hybridized to saturation with labelled probe D N A (1 mg cold D N A / m l hybridization buffer plus 100-500 ffl 32P-labelled probe, final spec. act. 350000-5000000 c p m / m g total DNA). Three different probes were used: pMF2 to probe for rDNA. and either pLH11 as a single-copy probe or pVK88 as a 'double copy' probe. The ratio of total pMF2 hybridized compared to either pVK88 or p L H I 1 hybridization was used to determine the number of individual r R N A cistrons present within the repetitive r D N A sequence. Table I shows the results of such an experiment for the AR33 parent, the freshly-generated D N O strain and the fifth and eleventh vegetative transfers of the D N O strain. The results indicate that the AR33 parent contains 225 repeat units while the D N O progeny have 440. The 340 repeat units observed in the fifth transfer is intermediate between AR33 and the original DNO, and indicates that a reduction of cistron number toward the original normal value

167 TABLE I ESTIMATES O F r D N A CISTRON N U M B E R F R O M S A T U R A T I O N . H Y B R I D I Z A T I O N O F r D N A A N D S I N G L E COPY D N A TO G E N O M I C D N A Source of genomic D N A

r D N A bound (#g)

Single Copy D N A bound (pVK88 t or pLHI 1 * (#g)

Estimated cistron number

T(VL ~ IVL)AR33

13.6±0.1 28.4--+0.7 21.6-+0.1 52.1 -+ 1.8 22.5-+0.1 25.9-+ 1.3

0.12±0.01 0.13--+0.01 0.10_+0.01 0.13-+0.01 0.13-+0.01 0.11 -+0.01

225 220 440 400 340 235

D N O first transfer D N O fifth transfer D N O eleventh transfer

t * t * t *

Genomic D N A from T ( V L ~ I V L ) A R 3 3 , D N O T No. l, D N O T No. 5, and D N O T No. I I was covalently b o u n d to diazobenzyloxymethyl cellulose as described in Methods. 32p-labeled D N A from p M F 2 ( r D N A repeat unit), pVK88 (qa.4 and qa-2, Giles, personal communication), or p L H l l (qa-2 only, Giles, personal communication), was used as the hybridization probe, Hybridization was conducted as described in the Methods section. The specific activity of the probes used in these experiments ranged from 0.3-5- 106 c p m / m g and the Cot value was 90. The ratio of single copy D N A hybridized to r D N A hybridized was used to estimate the number of r R N A cistrons in each genome. Results given are .X ± S.E., n = 8. Each sample was a 10 cm 2 section of the diazobenzyloxymethyl-cellulose paper.

has occurred. By the eleventh vegetative transfer, the rRNA cistron number has returned to the original parental value, within the limits of experimental error. The eleventh transfer has also regained some degree of fertility in sexual crosses, although fertility is below the wild type level. Discussion

This paper presents the results of a series of experiments involving a duphcation strain of N. crassa which contains two nucleolus organizer regions (the DNO strain), instead of the normal one. The following characteristics were observed in the DNO strain: 1. The DNO strain, while sterile, is otherwise physiologically normal and displays the same growth response as normal laboratory strains of Neurospora. 2. Comparisons of hybridization kinetics for cloned rDNA and genomic DNA from the parental AR33 strain and the DNO strain indicate that the DNO strain contains about twice as much rDNA as the parental AR33 strain per haploid genome. However, the amount of rDNA present in the DNO strain decreases with successive vegetative transfers as evidenced by the fact that the fifth transfer of the DNO strain displays hybridization

kinetics intermediate between those of the original DNO and the parental AR33. 3. Comparison of the relative amounts of single copy DNA and rDNA probes hybridized to genomic DNO and AR33 DNA at saturation indicates that the parental AR33 strain contains 225 copies of the rRNA repeat unit, while the DNO strain contains approx. 440 copies. The observed doubling of rRNA cistrons in association with a cytologically visible doubling of the nucleolus organizer is as expected, given the strong correlation between the nucleolus organizer and the chromosomal coding region for rRNA [4,19-24]. In fact, the observation of a 2-fold increase in rRNA cistrons confirms the conclusion of Perkins et al, [11], that the translocated segment in T(VL -~ IVL)AR33 contained all or most of the chromosomal rDNA. The gradual reduction in the number of rRNA cistrons observed in successive vegetative transfers of the DNO strain is very suggestive of a regulatory process analogous to magnification but working in the oppposite direction. Such a demagnification has been observed by Procunier and Tartof [25] in Drosophila strains possessing an excess of rDNA produced by unequal crossing over. Increases in rDNA resulting from repeated replication of some baseline rDNA segment have been

168

observed frequently [26-28]. The above experiments indicate that regulation of cistron number can occur in both directions, and that rRNA repeat units can be lost when in excess as well as added when deficient. Two basic mechanisms have been proposed for rDNA magnification and either may apply to demagnification. Tartof [12] and Ritossa [29] favor unequal crossing-over among sister chromatids as the magnification mechanism, and this would obviously apply to demagnification as well. Unequal crossing-over would allow the gradual loss of rRNA cistrons observed while retaining the caf-1 R gene, as observed in our experiments. Buonogiorni-Nardelli et al. [26], Brown and Blackler [30], and Endow [31 ] favor selective replication of rDNA sequences, either chromosomally or in extra-chromosomal rings, as the magnification mechanism. Endow and Glover [32] have further evidence that magnification proceeds differently in somatic and polytene cells. In polytene cells rDNA from one parent only is replicated repeatedly, while rDNA from both parents is replicated in somatic cells [32]. Regulation by complete loss of one set of parental genes does not appear to be the case in Neurospora, as the intermediate values for rRNA cistron number observed in successive vegetative transfers of the DNO strain is inconsistent with the rapid and step-wide decrease predicted by such an amplification method. This DNO strain and its vegetative progeny provide a useful tool for further studies of rDNA regulation. For example, determining whether rRNA genes from each parental genotype are lost equally or preferentially from one will assist in defining the cellular mechanisms of demagnification and magnification, and we are currently studying this problem. Understanding the regulation of rRNA cistron number has considerable intrinsic interest in the broad implications for understanding the regulation of nuclear-cytoplasmic interactions and the possible physiological significance of the extensive genetic redundancy observed in eukaryotic organisms.

Acknowledgements We thank David Perkins for his generous supply of strains and encouragement, we also thank Robert Metzenberg for supplying us with the plas-

mids pMF2 and pVK88 and Norman Giles and Laine Huiet for the pLH11 plasmid. This research was supported by grant GM26082 to P.J.R. from the National Institute of General Medical Sciences, N.I.H. Some equipment used in these experiments was provided by a Biomedical Research Support grant RR07168 from the N.I.H. and by Special Equipment Program grants CDP-8007801, CDP-8007790, PRM-791995 and TFI-8021641 from the National Science Foundation,

References I Brown, D.D. and Gurdon, J.B. (1965) Proc. Natl. Acad. Sci. U.S.A. 51, 139-146 2 McConkey, E.H. and Hopkins, J.W. (19641 Proc. Natl. Acad. Sci. U.S.A. 51, 1197-1204 3 Perr3, B.P.(19671 Prog. Nucl. Acid Res. Mol. Biol. 6, 219257 4 Ritossa, F.M. and Spiegelman, S. (19651 Proc. Natl. Acad. Sci. U,S.A. 53, 737-745 5 McClintock, B. (19341 Z. Zellforsch Mikrosk. Anat. 21. 294-328 6 Perkins, D.D. and Bar~, E.G. (19691 J. Hered. 60, 120-125 7 Free, S.J., Rice, P.W. and Metzenberg, R.L. (19791 J. Bacteriol. 137, 1219-1226 8 Cox, P.A. and Peden, K. (19791 Mol. Gen. Genet. 174, 17-24 9 Krumlauf, R. and Mar'zluf, G.A. (19791 Biochemistry 18, 3705 3713 10 Krumlauf, R. and Marzluf, G.A. (19801 J. Biol. Chem. 255 (3), 1138-1145 II Perkins, D D . , Raju, N.B. and Bar~', E.G. (19801 Chromosoma 76 (3), 255-275 12 Tartof, K.D. (1973) Genetics 73, 57-71 13 Vogel, H.J. (19641 Am. Nat. 98, 435-446 14 Wilcockson, J. (19751 Anal. Biochem. 66. 64-68 15 Alwine, J.C., Kemp, D.J. and Stark, G.R. (19771 Proc. Natl. Acad. Sci. U.S.A. 74, 5350 5354 16 Wahl, G. (19791 Proc. Natl. Acad. Sci. U.S.A. 76, 3683-3687 17 Alton, K.A., Hautala, J.A., Giles, N.H., Kushner. S.R. and Vapnek, D. (1978) Gene 4, 241-259 18 Rigby, P.W., Dieckman, M., Rhodes, C. and Berg, P. (19771 J. Mol. Biol. 113, 237-251 19 Phillips, R.L., Kleese. R.A. and Wang, S.S. (19711 Chromosoma 36, 79-88 20 Phillips, R.L., Welser. D.F., Kleese. R.A. and Wang, S. S. (1974) Genetics 77, 285-297 21 Ritossa, F.M., Atwood, K.C. and Spiegelman, S. (19661 Genetics 54, 819-834 22 Malva, C. Garguilo, G. and LaMantia, G. (19791 Mol. Gen. Genet. 172, 67-72 23 Kaback, D.B. and Halvorson, H.O. (19771 Proc. Natl. A c a d Sci. USA 1177-1180 24 Petes, T.D. (19791 J. Bacteriol. 138, 185-192 25 Procunier, J.P. and Tartof, K.D. (19761 Nature 263,255-257 26 Buongiorno-Nardelli, M., Amaldi, F. and Lava-Sanchez, P.A. (19721 Nature 238, 134-137

169 27 Phillips, R.L. and Phillips S.G. (1973) J. Cell Biol. 58, 54-63 28 Mohan, J. and Flavell, R.B. (1974) Genetics 76, 33-44 29 Ritossa, F.M. (1968) Proc. Natl. Acad. Sci. U.S.A. 59, 1124-1131

30 Brown, D.D. and Blackler, A.W. (1972) J. Mol. Biol. 63, 75-83 31 Endow, S.A. (1980) Cell 22, 149-155 32 Endow, S.A. and Glover, D.M. (1979) Cell 17, 597-605