Gene amplification mechanism for the hyperproduction of T4 bacteriophage gene 17 and 18 proteins

J. Mol. Biol. (1987) 195, 769-783

Gene Amplification Mechanism for the Hyperproduction of T4 Bacteriophage Gene I7 and 18 Proteins Duu Gong Wu and Lindsay W. Black Department of Biological Chemistry University of Maryland School of Medicine Baltimore, MD 21201, U.S.A. (Received

20 March

1986, and in revised form 9 September

1986)

Bacteriophage T4 mutants hyperproducing gene 17 protein (Hp17) have been isolated at high frequency by growing gene 17 amber mutants on ochre suppressor strains of Escherichia coli. Most mutants showed the co-hyperproduction of gene 18 protein, although one anomalous mutant hyperproduced a 60,000 M, partial polypeptide of gene 18. Hybridization of T4 late RNAs to cloned plasmid DNA containing genes 17, 18 or control T4 genes revealed that approximately five times more gene 17 mRNA and two to three times more gene 18 mRNA were synthesized in the Hp17 mutant infections. DNA-DNA hybridizations showed that Hp17 mutant DNA contained two to three times more copies of genes 17 and 18 than wild-type DNA. Southern blot analysis suggested that Hp17 mutants carry a direct tandem repeat of the gene 17-18 region, with variable copy number from one to at least six copies‘. Hyperproduction of gene 17 and 18 proteins appears therefore to result from gene amplification specific to the gene 17-18 region. In contrast to gene duplications reported in lambda and T4 phage, and numerous cases of gene amplification in bacteria, a similar or identical novel junctional fragment created by the amplification event was observed among 28 independent T4 Hp17 isolates; therefore, the mechanism giving gise to amplified sequences may involve specific sequences in this region of the T4 genome.

conditions no longer demand the presence of multiple gene copies, amplified sequencescan more easily be lost by recombination and segregation. Gene amplification has been widely reported in While the importance of gene amplification in both prokaryotes and eukaryotes (for reviews, see both bacteria and eukaryotic cells has been wellAnderson & Roth, 1977; Stark & Wahl, 1984; established, its mechanism of origin is still poorly Schimke, 1984; Long & Dawid, 1980). This increase understood. The molecular mechanism underlying in gene copy number can take place spontaneously specific instances of gene amplification could during development, as in amplification of the involve illegitimate recombination, replication or rRNA genes of Xenopus oocytes (Tartof, 1975) or other processes.In fact, it is likely that there are a the chorion genes of Drosophila (Spradling & Mahowald, 1980). Amplification can also occur number of distinct mechanisms that can give rise to amplification in different biological systems (Stark under environmental selection, as for example in such as methotrexate in & Wahl, 1984; Schimke, 1984). In bacteriophage, response to drugs mammalian cells (Alt et al., 1978; Kaufman et al., gene amplification has been studied relatively 1979) or to growth conditions in the selection of infrequently and only three casesin T4, lambda and Escherichia coli mutants carrying amplified lac P2 have been well characterized (Weil et al., 1965; Bellett et al., 1971; Chattoraj t Inman, 1974). genes (Albertini et al., 1980; Tlsty et al., 1984). The Bacteriophage T4 develops largely independently of frequent observation of gene amplification has suggested that it might represent an important its host, with autonomous and simpyified recombination and replication systems. For these reasons, form of regulation of gene expression as well as play a significant role in evolution. Under conditions phage T4 offers a distinct and possibly attractive model system for mechanistic studies of gene where a large and rapid accumulation of specific gene product is required, amplified genes provide a amplification (see Discussion). Here, we report the isolation in high frequency from amber mutants in flexible means of meeting such a demand without having to select a change in DNA sequence. When gene 17, Hp17 mutants capable of hyperproducing 769 1. Introduction

0022-2%36/87/120769-15

$03.00/O

0

1987 Academic

Press Inc. (London)

Ltd.

770

Lj.

G.

Wu

and

gpl7t and gp18 when grown in ochre suppressor strains of E. eoli. These mutants carry unstable multiple copies of the gene 17-18 region that result, in the hyperproduction of the corresponding mRNAs and proteins. The presence of an identical or similar junctional fragment in a large number of individual isolates suggests that the amplification event may occur frequently at a limited number of sequences.

2. Materials and Methods (a)

Materids

E. coli B40 su+T, B40 su+II. CA180 su+II. BE sum were used as permissive and non-permissive hosts for T4 amber mutants. CA165 sufB and CA254 su+E (Brenner & Beckwith, 1965) obtained from Barbara Bachmann of the Yale E. coli Genetic Stock Center are permissive host’s for T4 ochre mutants. The ~600 series recombinant plasmids used have been described by Mattson et al. (1977) and Seizer et al. (1978) and were kindly provided by Drs E. T. Young and T. Mattson. T4 bacteriophages are T4 wild-type, 16amN178I . used 16amN66, 17amNG178, 17amA465, 17amNGlWalt. 17amA465-alt. 18amE18 and alt (Horvitz, 1974; Goff, 1979). The phage used to prepare T4dC DNA is a pentuple mutant 42amN5556amE51-denA-denB-ale (Snyder et al., 1976) kindly supplied by E. Kutter. Hershey broth (H Broth), M9 and M9S medium have been described (Black & Ahmad-Zadeh. 1971). M9S.l contains 0.1 of BactoCasamino acids in M9S medium, Restrirtion enzymes E’coRV, Tag1 and NdeI were purchased from Bethesda Research Laboratories and New England Biolabs. [MethyZ-3H]thymidine (50 Ci/mmol), [5,6-3H]uridine (42 Ci/mmol) and 14C-labeled amino acid mixtures (1.78 mCi/mg) were purchased from ICN Radiochemicals. [a-32P]dATP and dTTP (0.8 Ci/mmol) were from New England Nuclear. An antiserum against gene 18 protein was kindly provided by Dr Fumio Arisaka of Hokkaido Universit~y, Japan. (b) Isolation

of gp17

hyperproducing

mutants

For selection, gene 17 amber mutants were combined with an alt mutation to eliminate the alt protein (Horvitz, 1974; Goff, 1979), which frequently co-migrated with gp17 on polyacrylamide gels. Two gene 17 amber mutants (17amNG178-alt and 17amA465-alt) containing a low frequency of wild-type revertants (< 10m6) were plated on ochre suppressor strains of E. co& (CA165 sufB or CA254 su+E), and incubated at 37°C overnight. Plaques formed by presumptive revertant phages were picked and characterized by their ability to grow on different hosts. Wild-type revertants were discarded, and revertants still carrying the original gene 17 amber mutation, but able to plate on the ochre suppressor strains, were grown into stocks in CA165. The quantity of gp17 synthesized by these mutants was rapidly screened by labeling with 14C-labeled amino acids in an amber permissive host (B40 su+I), followed by SDS/polyacrylamide gel electrophoresis. The independent isolates of t Abbreviations used: gp, gene product; T4dC, T4 deoxycytosine-containing DNA; m.o.i., multiplicity of infection; Hp17, gene 17 protein hyperproducing: TEMED, tetraethylmethylene diamine; HMC. hydroxymethylcytosine.

L.

W. Black

Hp17 mutants were prepared from separate stocks grown from single plaques of gene 17 amber-alt mutantas. (c) l’reparatim of “(‘-labeled amino acid-labelled lysates of T&infected E. wli nnd Sns/polyncrylnntidr gel electrophoresis E.

coli

H40

su’

1 or

BE:

sum

were

growr

to

4x

IO”

bacteria/ml in M9 medium. 20 pg I,-trgpt,ophan/ml was added and then the strains were infected at’ 30°C with T4 mutants at a m.o.i. of 8. T4 late proteins were labeled by addition of a 14(‘-labeled amino acid mixture to the infected cult,ure (1.5 pCi/ml) at 18 min after infection and incubation was continued for another 10 min. Infec%ed cells were centrifuged (8000 revs,/min at ~4”(’ for 10 min), resuspended in sample buffer, and subject’ed t’o HDS/polyacrylamide gel electrophoresis according to Laemmli (1970), with slight modifications. The final concentrations of the separation gel were as follows: 81;) (wI’~) acrylamide. 0.21 “/0 (w/v) N,N’- methylene-bisacrvlamidr, 0.15% (w/v) SDS, 0.03% (w/v) ammonium persulfatt,. 0.037590 (v/v) TEMED; 0.375 M-Tris . HCl (pH 8.8). The stacking gel contained 5?,, acrylamide. 0.25”,, N, N-methylene-bisacrylamide. O.l5yq] SDS. 0.1 “<, TEMED, 0.125 M-Tris. HCl (pH 6.8). Before loading onto the gel, the lysate was heated at 90°C for 3 min. After elect,rophoresis was completed, the gel was dried and autoradiography was carried out, for 2 to 6 da,vx using Kodak X-OMAT AR film. (d)

Isolation

of DiVA, and preparatzwn DNA -bound jilters

oj’

Bacteriophage T4 DNA was isolated from cesium chloride gradient-purified phage particles by gentle extraction with phenol and dialyzed against TE buffer (0.01 M-Tris.HCl (pH 8.0). 0.001 M-EDTA) overnight. Plasmid DNA was isolated bv an alkaline extraction and polyethylene glycol precipitation procedure (Birnboim & Doly. 1979). followed by RNase digestion. Bio-Gel A5m column chromatography. and precipitation with ethanol (Just et nl.. 198.3). Filt,ers csont,aining immobilized phage (11’plasmid 1)X;\ were prepared essentially as described by Young rt al. (1980). DNA (5 to 10 pg/ml in 20 ml of 0.1 br-NaOH) was heated for 10 min at, 67°C‘ to hydrolyze ribonucsleotides in the plasmid DNA and to denature the DNA. Follow ing the addition of 20 ml of neutralizing solution (10 ml of 1 M-HCl and 5 ml of 1 M-Tris. H(‘1 (pH 8.0) were mixed with 100 ml of water). the denatured DNA was boiled in a waterbath for 5 min and quickly chilled by adding 140 ml of cold water. Prior to loading on to filters. 20 ml of cold 20 x SSC (SS(’ ix 0.15 M-NaCl. 0.015 M-trisodium citrate) and 2 ml of 3 M-MgCl, were added to the solution. A volume of DNA solution containing approximately 20 pg of DNA was t,hen slowly loaded on to a 25 mm nitrocellulose filter mounted on a chilled filtration unit under mild vacuum. After washing twice with 2 x SSC. the filters were air-dried. baked for 2 h at 80°C in a vacuum oven, and cut) into 9 mm discs with a cork borer before hybridization. T4 DNA filters were prepared by the same procedure after t,he DNA was fragmented using a Biosonix TV sonicator. (e)

Prepa.ration and

of RNA,

radioactively and

labeled hybridization

I’d

II&A

E. coli (f, ml) grown to 3 x 10’ cells/ml in MY medium supplemented with 5 pg thymidine/ml were pelleted and

Amplijication

of Bacteriophage

resuspended in 5 ml of M9 medium containing 250 pg 2-deoxyadenosine/ml, thymidine/ml, 5 I@ 25 pg L-tryptophan/ml, and 5 PCi of [methyZ-3H]thymidine/ml. After incubation for another 30 min at 37”C, the culture was infected (m.o.i. 5) and, after 90 min, the labeled phages were purified on a cesium chloride step gradient. Radioactive T4 DNA was extracted with phenol and sonicated for 25 s by using a Biosonix IV sonicator. The size of sonicated DNA under these conditions is approx. 400 base-pairs by comparison with DNA fragments of known size on an agarose gel. Following denaturation at 95°C for 3 min, DNA containing 1.0 x lo5 cts/min was added into 1 ml of hybridization solution (6 x SSC, 5 x Denhart solution, 0.5% SDS, 0.01 M-EDTA, 100 pg denatured calf thymus DNA/ml, and 50% formamide) containing a prehybridized filter in a capped vial and incubated at 42°C for 48 h with shaking. After hybridization. filters were washed once with 2 x SSC, 0.57; SDS and then 2 x SW, 0.1 y0 SDS for 15 min each at’ room temperature, followed by incubation with 0.1 x SSC and 0.5% SDS at 37°C for 1 h with shaking. The filters were then air dried and the retained radioactivity was counted in a scintillation counter. %labeled T4 late RNA was prepared as described by Young et a,Z. (1980). E. coli (5 ml) grown at 37°C to 2 x 108 cells/ml in M9S medium containing 10 pg uridine/ml were pelleted, washed once with M9S, and resuspended in 0.25 vol. M9S.l containing 50 pg z-tryptophan/ml. The concentrated culture was then infected at 30°C with phages suspended in an equal volume of M9S.l (m.o.i. 5). [3H]uridine was added to the culture to a final concentration of 50 to 100 @Zi/ml at 15 min after infection. Incorporation was stopped at 30 min by pouring the infected culture into 2 vol. cold Tris. HCljKCl buffer (0.1 M each, pH 7.0) containing 10 pg uridine/ml and 2 mM-KCN. Radioactive RNA was extracted by the rapid lysis, phenol/DNase procedure (Young & Van Houwe, 1970). RNA-DNA hybridization was carried out in a 2 x SSC solution containing 50% formamide. Duplicate filters for each sample were placed in screw-cap vials, containing 1 ml of hybridization solution to which approximately 5.0 x lo5 cts/min of radioactive RNA was added. After incubation at 40°C for 3 days, the filters were rinsed and then washed with 2 xSSC at room temperature for 15 min. After 4 to 5 washes, the filters were treated with pancreatic RNase (50 pg/ml) at room temperature for 30 min, dried. and filter-bound radioactivity was counted. The amount of excess cloned plasmid DNA necessary t’o ensure the exhaustive hybridization of radioactive RNA was used as determined by Young et al. (1980). (f) Restriction

enzyme digestion and agarose gel eleetrophoresis

Conditions for restriction enzyme digestion of T4 glucosylated HMC-DNA are as follows: For EcoRV, 1.5 pg of DNA was incubated with 15 units of enzyme overnight at 37°C for 16 h in a volume of 20 ~1 containing 50 mM-Tris.HCl (pH 9.0), 5% (v/v) glycerol, 10 mM-MgCl,, 10 mw-dithothreitol, and 10 pg of bovine serum albumin/ml. NdeI digestion was carried out with 10 units of enzyme in 20 pl of reaction mixture containing 50 mi%-Tris. HCl (pH 8+0), 10 mM-MgCl,, 50 mM-NaCl at 37°C overnight. Both EcoRV and NdeI-digested DNAs were electrophoresed on a 0.7% agarose gel in 0.004 M-Tris-borate (pH 8.0) containing 0.001 M-EDTA. TaqI digestion was carried out in a 25-~1

T4 Genes 17 and 18

771

volume containing 1.5 pg of DNA, 10 units of enzyme, 50 mM-Tris.HCl (pH 8.0), 50 mM-NaCl, 10 mM-MgCl,. The reaction mixture was placed in a small Eppendorf tube, covered with 200 ~1 of paraffin oil, and incubated for 16 h at 65°C. Digested DNA was then electrophoresed on a ly/, agarose gel in 0.004 M-Tris-acetate (pH 7.8) containing 0.001 M-EDTA. (g) ~Viclz translation, Southern hybridization

blotting. and

Plasmid DNA containing T4 inserts was labeled by nick translation (Rigby et al.? 1977) with [a-32P]dATP or dTTP. Labeled DNA was separated from unincorporated nucleotide on a Sephadex G-100 column. Sfter gel blotting (Southern, 1975), hybridization was performed in a sealed plastic bag under conditions described for DNADNA filter hybridization. After hybridization, filter papers were washed, dried, and exposed to a Kodak X-OMAT AR film with a DuPont Cronex enhancer screen at -70°C. (h) Immunoblotting Protein samples were electrophoresed on a polyacrylamide gel and electrophoretically transferred to a nitrocellulose sheet essentially as described by Towbin et al. (1979). After transfer, proteins of interest were detected immunologically with antibody. followed by subsequent incubation with goat anti-rabbit immunoglobulin G-peroxidase conjugate. A positive interaction was visualized by using the substrate 4-chloro-1-naphthol (Howkes et al., 1982).

3. Results (a) Isolation

of mutants

hyperproducing

gp17

Bacteriophage T4 gp17 is required for DNA packaging and links DNA to the DNA entrance vertex (gp20) of the T4 prohead (Laemmli & Favre, 1973; Hsiao & Black, 1977; Black, 1981). Although required “stoichiometrically” (Snustad, 1970), gp17 is synthesized in very low concentration and its identification as a 68,000 M, protein by isotope labeling, SDS/polyacrylamide gel electrophoresis,

and autoradiography is disputed (Vanderslice & Yegian, 1974; Wunderli et al., 1977). The adjacent genes are 16. closely associated with gene 17 in packaging, whose 20,000 M, low-concentration product has been identified (V. B. Rao, personal communication),

and

gene

18, the structural

gene

for the major tail sheath subunit (70,000 M,), which is synthesized in abundance in T4-infected bacteria. Genes 16, 17 and 18 are “late” genes, whose

expression depends upon T4 replication and gene 55 protein Two

modification of E. coli RNA polymerase. E. coli strains carrying ochre suppressors

specifying the insertion of a glutamine at the ochre nonsense

codon

have been reported

to suppress

T4

amber mutants with a low efficiency of approximately 4 to 8% (Brenner & Beckwith, 1965). To select for mutants capable of hyperproducing gp17, we searched for mutations that overcome the effect of the low level of gp17 resulting

from

this

inefficient

suppression

of the 17

D. G. Wu and L. W. Black

772

amber mutation in these ochre suppressor strains. When* two gene 17 amber mutants (amNG178-alt and amA465-alt) were plated on these two ochre suppressor strains, most phages were unable to form plaques, although the phage were able to grow on glutamine-inserting amber suppressor strains (B40 su+II and CA180 su+II). However, when a high concentration of phages was plated, minute plaques appeared with a frequency greater than that of the wild-type revertants. These revertants were first tested for their ability to grow in various E. coli hosts containing or lacking suppressor genes. Phages that grew on the ochre strains while still retaining the original amber mutation in gene 17 were then subjected to a rapid screening of the gpl7

I

2

3

4

5

synthesis by 14Clabeled amino acid labeling and SDS/polyacrylamide gel electrophoresis. Revertants showing hyperproduction of gp17 in an amber permissive host (B40 su+I) were designated gene 77 protein hyperproducing mutants (Hpl7). Four Hp mutants, Hp17-5. -47. -62 and -64, derived from two gene 17 amber mutants at a frequency of 10m5 to 10e6 (approx. 5 to 10 times higher than the frequency of wild-type revertant) were initially characterized. Figure 1 shows an autoradiograph of “C-labeled T4 late proteins synthesized in Hpl’l-infected amber permissive (Fig. l(a)) and non-permissive (Fig. l(b)) hosts. Compared with the parental amber mutant (NG178-alt) in which the gp17 was invisible on this

I2

345

9~23’ 9~20

(a)

Figure 1.

(b)

SDS/polyacrylamide gel electrophoresis of T4 late proteins. Samples of 14(J-labeled proteins were prepared from (a) amber suppressor E. coli B40 su+I and (b) non-suppressing E. coli BE su- infected with 17amNG178-alt (lane l), Hp17-5 (lane Z), Hp17-47 (lane 3), Hp17-62 (lane 4), and Hp17-64 (lane 5) as described in Materials and Methods. Gel (a) was a 16 cm x 15 cm, SDS/s% polyacrylamide gel while gel (b) was 16 cm x 20 cm. Equal amounts of radioactive protein were electrophoresed, and the gel was dried and exposed to an X-ray film for 2 days. The indicated T4 gene products were identified according to the method of Vanderslice & Yegian (1974). as well as by comparison with gels containing labeled T4 late proteins resulting from infections with appropriate T4 amber mutants. gp23* represents the cleaved form of the major T4 capsid protein. K, lo’&

AmplQication

of Bacteriophage

autoradiograph (Fig. l(a), lane l), the hyperproduction of gp17 was evident in the lysates from Hpl7-infected amber permissive E. coli (lanes 2 to 5). From the pattern of T4 late proteins, two types of Hp17 mutants were distinguished. The concentration of gp17 synthesized by Hp17-5 (Fig. l(a), lane 2), an apparently distinctive Hp17 mutant, is higher than that of the parental amber mutants (lane 1) but lower than that of Hp17-47, -62 and -64 (lanes 3 and 5). Furthermore, a clear distinction between the Hp17-5 and the other Hp17 mutants was evident from the co-hyperproduction of a second protein in each type. While a marked increase of gp18 was observed in the infections with Hp17-47, -62 and -64 (Fig. l(a), lanes 3 to 5), an unknown 60,000 M, protein was found to be hyperproduced by Hp17-5 (lane 2). Hp17 infections of the non-permissive host (BE su- ) showed hyperproduction of both the 60,000 M, or gp18 proteins, but synthesis of gp17 was totally absent, apparently due to the presence of the gene I?’ amber mutation (Fig. l(b), lanes 2 to 5). Aside from the hyperproduction of gp17, gpl8 and the 60,000V~, -protein, synthei& of Tther 1234

-m ame

01t+

20

*

II)

-w34

*

T4 Genes 17 and 18

773

proteins appeared to be unchanged in the hyperproducing mutants. The relative concentrations of gp17 and gp18 synthesized by two representative Hp17 mutants were quantified by densitometric scanning of an autoradiograph containing well-separated protein bands (Fig. 2(a)). The results showed a more than fivefold and two- to threefold increase in the amount of gp17 and gp18, respectively, in Hp17-47 over wild-type phage (Fig. 2(b)). Hpl7-5 produced approximately as much gp17 as did wild type, which still represented a level higher than that of the amber parental mutant. The increase of the 60,000 M, protein could not be determined because there was no counterpart in wild-type and gene 17 amber mutant infections. Table 1 shows the growth characteristics of the Hp17 mutants, the parental gene 27 amber mutants and wild-type T4 on various hosts. The presence of the original gene I7 amber mutation in the Hp17 mutants was demonstrated by their low plating efficiencies on a non-permissive host (BE su- , Table 1A) and by complementation and recombination tests (data not shown). The plating

3 I

QPl6 QPl7 QP20 6OK

GP34

QP18

9P17

(b)

Figure 2. Relative concentration of gp17 and gp18 synthesized by wild-type T4, 17amNG17%alt, Hp17-5 and Hp17-47. (a) Autoradiograph of 14C-labeled T4 late proteins prepared from T4-infected amber suppressor E. coli B40 ~‘1, following polyacrylamide gel electrophoresis. (b) Autoradiograph (a) was scanned with a densitometer and the peak areas on the chart corresponding to proteins of interest were measured by weighing, and changed into the relative protein concentrations using gp34 as reference. The alt protein in the TP wild-type-infected lysate co-migrated with gp20 on this gel. K, lo3 M,.

D. G. Wu and L. W. Blaclc

774

Table 1 Growth A.

Efieiency

properties

of Hp17

mutants

of plating Host

Phage

B40 su+1

T4’D 17amA465-alt 17amNG178-ah Hpl7-5 Hp17-47 Hp17-62 Hp17-64

1 1 1 1 1 1 1

HE su0.98 I.0 x 1.1 x 1.0 x 1.3 x 4-7 x 5.7 x

T4+D 17amNG178-alt Hpl7-5 Hp17-47

su+ls

0.9 1.2 x 1W5 5.1 x 1W6 0.42 0.48 0.48 0.44

1om6 1om6 10-3 10Fh 1o-h 1o-6

(:A165 su+lS

IS40 su+1 Phage

CA165

0.1t

0.1t

1 0.92 0.93-1.35 1.05-l .38

I < 1o-4 0.28-0.37 0.28-0.36

5t 1 < 10F4 0.44-0.54 0.4440.58

Phages of appropriate dilution were plated on different E. coli hosts. The efficiency of plating to that on B40 su+T. Hpl7-5 was originally isolated from 17amA465-ah on ochre suppressor strain CA254 su+E, while Hp17-47. -62 and -64 were selected from 17amNG178-ah on strain CA165 su+R. However, all Hp17 mutants were subsequently grown in CA165 for all the experiments. The measurements of burst sizes were carried out 3 times for each phage at 37°C using H40 su+I (m.o.i. 0.1) and CA165 (m.o.i. 0.1 and 5) grown in H broth. t ~Multiplicitg of infection.

efficiencies of the Hp17 mutants on ochre suppressor E. coli increased from lo-’ to 1O-6 of the parental amber mutants to approximately 0.4 to 0.5. The burst sizes of the Hp17 mutants in the ochre suppressor strain were approximately 30% (m.o.i. 0.1) to 60% (m.o.i. 5) of the wild-type infection, while the relative burst size for the parental amber mutant was less than O.Oly& (Table 1B). (b) The hyperproduction of gp17 and gp18 is the result of elevated concentrations of their mRNAs To determine whether the hyperproduction of gp17 was a result of a mutation affecting transcription or translation, we first examined the concentration of gene 17 and 18 mRNAs in an Hpl7-infected amber suppressor host (B40 su+I). The availability of cloned plasmid DNA containing T4 inserts (Fig. 3) allowed us to perform the measurements directly by RNA-DNA hybridization. The transcription of both genes 17 and 18 has been reported to begin at seven minutes at 30°C and continue at an increasing rate until at least 20 minutes after infection (Young et al., 1980). Therefore, T4 late RNAs labeled with [jH]uridine between 15 and 30 minutes were extracted and hybridized to nitrocellulose filter-bound DNAs. Of

all the plasmid DNAs tested. the amount of mRNA hybridized t,o genes 12 to 16 (p659), 27 t,o 23 (~652) or 24 (~654) remained relatively constant among all the phages (Fig. 4). However, hybridizations with gene 17 and 18 internal fragments (~666 and ~655) showed that approximately five times and t’wo to three times the wild-type level of gene 17 and IX mRNAs accumulated in Hp17-47 infections. In Hp17-5 infection, gene 17 and gene IX mRNA levels were twice the wild-t)ype value (Fig. 4). The increase in the production of gene 17 and IX mRNAs by the Hp17 mutants is in general agreement) with the level of hyperproduction of the proteins, given the possible drffering efficiencies of suppression at, the NG178 and A465 amber codons. Therefore, the hybridizations supported the conclusion that hyperproduction of gene 17 and 1X proteins by Hp17 mutant’s reflect’s an increase in their mRNA levels. (c) The elevation of genr 17 and 18 &LV/ls dur to gene amplification

Ls

Initially we speculated t,hat, up-promoter mut,ations found frequently in phage and bacteria (Calos. 197X; Meyer ef trl., 1975) might be responsible for the elevated level of gene 27 and IX mRNAs. However, unlike up-promo&r mutants, hyperproduction of gp17 and t,he abilit,y to grow in an ochre suppressor host observed in Hp17 mutant,s appeared to be very unstable. The instability was evident in a coupled loss of both gp17 and gpl X hyperproduction when Hpl7 mutants were grown again on an amber suppressor host (B40 su+ I). Such instability, and bhe co-hyperproduction of the product, of a nearby gene. two common rharacbtrristics of gene amplification observed in both prokaryotes and eukaryotes, suggested 1hc existence of multiple copies of genes 17 a.nd IS in the Hp17 mutants. Therefore, L)SA~-D&A hybridization was performed to quantify the content of genes 27 and 18 in the Hpl7 mut,ants. 3H-labeled Hp 17 and wild-type I)NAx were prepared. and hybridized to an excess amount of T4 cloned plasmid DNAs immobilized on to nitro cellulose fibers. Figure 5 shows the results of the hybridizations. We found that t,he amount of DNA hybridized to plasmids containing T4 genes 21 to 23 (~652) and 24 (1’654) remained relatively the same in all the DNA species. However. not)iceable differences were observed when c~loned I)N As bearing t)hc gent 17-f8 region were used. .?is compared t,o wild type or the parental amhet mutant (17amXG178-ait,; data not shown). two to three times more DNA from both the HP17-5 and -47 mutants hybridized t,o the plasmid I)NAs c*ontaining an individual internal gem 17 (~666)~ gene 17-78 junctional (~662). or internal gene IS (~655) fragment (see Fig. 3). Although a greater amount of Hp17 DNA wa,s also hybridized to the gene 7fi and gene 12-16 clones (pRl6 a.nd p659), which share an identical end point, it was later found t’hat t,his rna,y result from a small portion of

Amplification

of Bacteriophage

I6

Genes

I6

17

19

T4 Genes 17 and 18

20

l

Y

pRl6 -

Probes

-

-

23

24 . ........I....

pCBR2

P666 Y

c

21 22 ’

775

~662 ~655

. ~652 -

,

a p659

~654

I kb

Figure 3. Size and genetic location of cloned T4 fragments. The location, size and orientation of T4 genes in the upper circular map weretaken from Young et al. (1980).The co-ordinatesin the inner circle correspondingto the bases( x 103) of T4 DNA are according to Kutter & Rueger (1983). All of the 600 series plasmids are EcoRI fragments cloned into either pBR322 or pBR313 vectors by Mattson et al. (1977). pRl6 containing gene 16 and part of gene 17 (approx. 100 base-pairs) and pCBR2 containing a 6900 base insert of genes 16 to 20 were constructed in this laboratory by V. B. Rao (Rao & Black, unpublished results). kb, lo3 base-pairs. gene 17 sequences (approx. 100 base-pairs) present in these two clones. In a subsequent mapping experiment with restriction enzymes, we concluded that an intact gene 16 was not amplified (data not shown). Very similar results were observed when DNAs of other Hp17 mutants, Hp17-62 and -64, were used (data not shown). At the same time, multiple copies of gene 17 and 18 sequenceswere no longer detected, when Hp17 phages were reverted to

a

parental-type

amber

mutant

on

an amber

suppressor E. coli (R40 su+I; data not shown). These measurements clearly suggest that Hp17 carries multiple copies of the gene 17 and 18 region, and that this is the primary cause of the hyperproduction of the corresponding mRNAs and proteins. (d) IdentiJication of the ampli$ed sequences by restriction enzyme digestion, Southern blotting and hybridization Although the filter hybridization experiment suggested that the Hp17 mutants may carry two to three copies of both genes 17 and 18, the sizes and the arrangement of the amplified sequences were unknown. Consequently, we analyzed the DNA structure of Hp17 mutants by Southern blot

analysis. In the 166,000 base T4 DNA molecule, cytosine is replaced by glucosylated hydroxylmethylcytosine (glu-HMC), which makes T4 DNA resistant

to most

restriction

enzymes.

Exceptions

include TaqI, EcoRV, NdeI and AhaIII. From a detailed restriction map of T4dC DNA constructed by Kutter t Rueger (1983), cleavage sites for NdeI and EcoRV were found to be located outside the gene 16 and 18 region. Therefore, these two enzymes were used to analyze the amplified sequencescarried by the Hp17 mutants. Figure 6(a) shows the patterns of NdeI digests of T4 DNA stained with ethidium bromide after electrophoresis on an agarose gel. While a group of identical fragments smaller than 9400 bases were observed among all the DNA species, an array of fragments larger than 9400 baseswas visible only in the DNA digests from Hp17-5 and -47 (Fig. 6(a), lanes 3 and 4) but not from wild type (lane 1) or the parental amber mutant (lane 2). These fragments were electrophoresed further to obtain a better separation on the gel, as seen in Figure 6(b), and were then transferred to nitrocellulose sheets, followed by hybridization with a probe containing gene 17 sequences (~666). A strong hybridization between the gene 17 probe and the multiple bands observed above 9400 bases in the Hp17 DNA

776

D. G. Wu and L. W. Black

T4 cloned

plasmid

DNA

Figure 4. Hybridization of 3H-labeled T4 late RNA to cloned plasmid DNA containing T4 inserts. [3H]RNA~ labeled from 15 to 30 min after infection of E. coli B40 sucI by wild-type T4 and 2 Hp17 mutants were purified and hybridized to cloned plasmid DNA containing T4 inserts as described for Fig. 3. Duplicate samples containing approximately 5.0 x lo5 cts/min each were used for hybridization. The results are expressed as a percentage of the radioactivity hybridizing to the total T4 DNA. The radioactivity hybridized to pBR322 vector DNA alone was subtracted as background. The difference in the bound radioactivities between 2 individual filters was smaller than $90/ of the average value in all the samples. kb, lo3 base-pairs. digests (Fig. 6(b), lanes 7 and 8) indicated that they were fragments carrying amplified gene 17 sequences. In addition, a doublet of bands of approximately 6600 bases, which appears to represent the normal NdeI fragments containing gene 17 sequences, was observed among all the DNA species (Fig. 6(b), lanes 5 to 8). Each one of the amplified fragments in Hp17 DNA digests (Fig. 6(b), lanes 7 and 8) appeared to exist also as a doublet commensurate with this 6600 base-pair. The formation of such a doublet is probably due to the variable cleavage of T4 HMC-DNA caused by the modification of cytosine, since digestion of T4dC DNA revealed only one corresponding gene 17 fragment of approximately 6300 bases (data not shown). From a detailed NdeI restriction map of T4dC DNA (Kutter & Rueger, 1985), two NdeI cleavage sites are located very close to one end of the 6300 base fragment (approx. 450 and 650 bases away). The modified bases surrounding this 6300 base site may cause it to be resistant to NdeI cleavage. Cutting HMC-DNA at the neighboring site 450 basesdownstream will therefore generate a doublet of 6300 and 6800 bases. Co-migration of these fragments with DNA from other portions of the T4 genome (e.g. the 6800 base fragment is expected

to overlap

a 6800

base

NdeI

fragment

from the gene 39 region) would explain the apparently higher intensity of this ethidiumbromide-stained doublet relative to that of amplified sequenceson the agarose gel but not on the autoradiograph (Fig. 6(b)). An estimation of the sizes of the Hpl’l-specific amplified sequences (Fig. 6(b), lanes 7 and 8) suggested that each fragment was probably derived from the stepwise addition of the same amplified unit to the 6600 base doublet of normal gene 27 fragment. The lengths of the amplified units were approximately 3800 basesand 4300 basesfor Hp 17-5 and -47, respectively. The presence of the 6600 base gene 17 monomer in all the DNA species also indicated that Hp17 mutants, although repeatedly purified from single plaques, actually exist, as a mixed population of phages carrying monomer, dimer, trimer, tetramer and even higher copy numbers of gene I7 (see Discussion). We estimated from the relative intensity of individual bands (Fig. 6(b), lanes 7 and 8) that phages carrying a single copy of gene 17 account for approximately 28 to 30% of the total population of Hp17-5 and -47 mutants, while 27 to 34% and 31 to 39% have two copies and more than two copies of gene 17, respectively. Southern blot analysis of EeoRV-digested Hp17

Amplijkation

of Bacteriophage

T4 Genes 17 and 18

777

*O-

lid

T4+ D Hpl7-5 Hpl7-

*5 -

47

,O-

,5 -

O-

5-

p~~6

(916)

pSSS(gl7)

~662

(g17-la)

p659 ~655

(g 18)

T4 cloned

plosmid

(g 12-K)

~654

~652

(g24)

(g21-23)

DNA

Figure 5. Hybridization of 3H-labeled DNAs of wild-type T4 and Hp17 to cloned plasmid DNA containing T4 inserts. Preparation of radioactive phage DNA was carried out with [3H]thymidine in B40 su+I and CA165 su+B for wild-type phage and the Hp17 mutant, respectively. DNA isolated from phage particles by extraction with phenol was sonicated, denatured, and hybridized to filter-bound plasmid DNA containing T4 inserts as described in Materials and Methods. The size and genomic location of individual clones used are as shown in Fig. 3. Duplicate filters were used for each sample and the average retained radioactivity of 2 filters was expressed as the percentage of total input radioactivity. The difference in the retained radioactivity between 2 filters was less than f 8% of the average value in all samples.

DNAs also showed a result very similar to the NdeI digestion. In addition to a set of normal EcoRV fragments being hybridized by a gene 17 probe in all the DNA digests, extra bands of amplified sequences were detected only in the two Hp17 DNAs but not in the wild type and parental amber DNAs (data not shown). To confirm that gene 18 is also included in the amplification sequence, we have probed both NdeI and EcoRV-digested Hp17 DNAs by a gene 18 probe (~655). The same result shown by the gene 17 probe was observed, while no amplified fragments were detected by a gene 24 probe (~654: data not shown). Similarly, multiple gene 17 fragments generated by the variable cleavage of HMC-DNA also appeared in the EcoRV digestion, while the T4dC DNA yielded only one fragment after hybridization. (e) The 60,000 M, Hp17-5 protein is a partial product of gene 18 DNA-DNA

filter

hybridization

(Fig.

5) showed

that the amount of DNA hybridized to a clone containing the gene 17 and 18 junctional fragment (~662) and an internal fragment of gene 18 (~655) was approximately equal in Hp17-5 and -47.

However, the concentration of gp18 synthesized by Hp17-5 (Fig. l(a), lane 2) was far less than that of Hp17-47 (lane 3). We supposed that the 60,000 M, protein synthesized by Hp17-5 was possibly a partial product of gene 18. To study this possibility, the origin of the 60,000 M, protein was determined immunologically using antibody against gp18. 14C-labeled T4 late proteins were separated on an SDS/polyacrylamide gel (Fig. 7(a)) and immunoblotted with antibody, as is shown in Figure 7(b). In addition to the normal gene 18 protein visualized among all the samples, an extra protein at the position corresponding to that of the 60,000 M, protein on the gel was identified in the lysates from amber suppressor and non-suppressing E. coli infected by Hp17-5 (Fig. 7(b), lanes 3 and 7). This result strongly suggests that the 60,000 M, protein is a partial polypeptide of gene 18. Given the fact that Hp17-5 appears to carry a shorter amplified repeat sequence (see section (d), above), it is likely that the 60,000 M, protein was derived from a truncated gene 18 created by the fusion of amplified units within gene 18. The presence of both normal gp18 and 60,000 Mr protein in the Hp17-5 infected lysate (Fig. l(a), lane 2) also indicated that the Hp17-5 mutant carries an intact as well as a

778

D. G. Wu and L. W. Black I

2

3

4

5

6

7

8

kb

23.

I

9.4

6.6

4.4

2.3

(a)

(b)

Figure 6. Agarose gel electrophoresis of LV’deI-digested wild-type T4 and Hp17 DPI’As. and hybridization with 32P-labeled ~666 (gene 17) probe. NdeI-digested wild-type T4 ((a) lane 1: (b) lanes 1 and 5), 17amh’G178-ak ((a) lane 2: (b) lanes 2 and B), Hp17-5 ((a) lane 3; (b) lanes 3 and 7); and Hp17-47 ((a) 1ane 4; (b) lanes 4 and 8) DIVAS were electrophoresed on a 0.7 y. agarose gel. (a) Ethidium bromide-stained agarose gel showing all NdeI fragments. (b) High molecular weight fragments of gel (a) after prolonged electrophoresis. Some low molecular weight fragments containing no gene 16-18 region were run off the gel. (c) Hybridization pattern of (b) with 32P-labeled ~666 (int’ernal gene 17). truncated gene 18 in the same DNA molecule. The tendency of Hp17-5 to accumulate wild-type revertants (Table 1A) may reflect the fact that the existence of a truncated gene 18 protein interferes with the growth of phages, even though normal gp18 is produced, and gp17 is synthesized in adequate levels. (f) Identification of a novel junctional fragment created by gene ampli$cation To determine the arrangement of the amplified sequences in the Hp17 mutants, we identified the novel junctional fragment by Southern blot analysis. Hp17 DNAs were digested with TaqT restriction enzyme, which makes multiple cuts within the gene 17-18 region, and probed with a clone containing genes 16 to 20 (pCBR2; Fig. 3). Extensive cleavage by TaqI makes visual identification of the junctional fragment on the agarose gel impossible (Fig. 8(a), lanes 1 to 4).

However, upon hybridization with the probe, in addition to a set, of bands corresponding to the normal TaqI fragments of the gene 16-20 region (Fig. 8(b), lanes 1 to 4), a 1580 base and a 930 base extra fragment were detected only in the DNA digests of Hp17-5 and Hp17-47, respectiveIS (Fig. 8(b), lanes 1 to 4). These two extra Tag1 fragments appear t’o contain the novel junctional sequencescreated by the fusion of the 5’ end of a second copy of the amplified sequence to the 3’ end of the first copy. The single junctional fragment observed in both Hp17 mutants strongly suggests t’hat the amplified sequencesof heterogeneous length seen in Figure 6(b) are arranged as a direct tandem repeat, becausean inverted or a randomly dispersed one would generate more than one junct,ional fragment by restriction enzyme digestion. On the basis of the genes 16 to 19 sequence data (G. Mosig personal communication: F. Arisaka. personal communication), most of the ToqI fragments of T4 HMC-DNA corresponding t-o those of

Ampl@cation

of Bacteriophage

12345676

I

T4 Gen.es 17 and 18

779

2345676

9Pl6 60K

(a)

(b)

Figure 7. Sl)S/polyacrylamide gel electrophoresis and immunoblotting of the 60,000 M, protein synthesized from 14C-labeled T4-infected by Hp17-5. Two sets of samples containing equal amounts of radioactivity E. coli lysates were electrophoresed on an 8% gel. (a) After electrophoresis, the 1st part of the gel contaiuing 1 set of samples was dried and exposed to the X-ray film. (b) The 2nd set of samples on the remaining portion of the gel then was transferred electrophoretically to nitrocellulose paper, and immunoblotted with antibody against gp18. Samples on (a) and (b) were as follow-s: 1, 17amNG178-alt; 2, 17amA465-alt; 3, Hp17-5; 4, Hp17-47; 5, Hp17-62; 6, Hp17-64; 7. Hp17-5; 8, Hp17-47. An amber suppressor host B40 su+I was used to prepare the lysates in lanes 1 to 6. while a non-suppressing host BE su- was usedfor samplesin lanes7 and 8. K, 10’ M,.

except (genes 16 to 20) clones were identified fragments x and y (Fig. 8(b)), which are probably from the gene 20 region. The concentrations of these fragments were then compared by their

pCBR

relative

intensities

to fragment

c on a properly

exposed autoradiograph (Fig. 8(b) was overexposed). Bn approximately twofold reduction in the relative intensities of both fragments x and y was observed in Hp17-5 and -47 as compared with wild-type amplified

T4 DNA. sequences

It of

was concluded Hp17 mutants

that the at least

include fragments b, c and d of genes 17 and 18 but not fragments x and y. We were unable to determine the relative intensities of the other gene 16 and 19 fragments due to their small sizes and co-migration. (g) Comparison of the extra junctional fragments from independently isolated Hp17 mutants We have examined the Tag1 junctional fragments in 29 independent isolates of Hp17 mutants from the two gene 17 amber mutants (amNG178-alt and amA465-alt)

by Southern

blot

analysis

using

the

gene 16-20 probe (pCBR2). Figure 9 shows the hybridization pattern of DNAs from T4dC, a gene

17 amber-alt mutant, the four initial Hp.17 mutants (Fig. 9(a), lanes 1 to 6) and 15 representative new isolates (Fig. 9(b), lanes 1 to 15). Hybridization with this probe showed that Tag1 fragments were apparently identical between T4dC DNA and gene 17 amber mutant DNA. The slight downshift in the pattern of T4dC DNA is apparently due to the unmodified cyti%ine, which replaces the usual glucosylated HMC in normal T4 DNA, and which gives rise to a slightly higher mobility fragment (Fig. 9(a), lane 1). Among a total of 29 such individual isolates, 26 have shown an apparently identical junctional fragment of approximately 930 bases (Fig. 9(a) and (b); 10 not shown), while only two, including Hp17-62 and one new isolate (Fig. 9(a), lane 5; (b), lane.‘l3), displayed basically the same extra fragment with very slight size variation

(less than

50 base-pairs).

Therefore,

the

930 base-type mutant is the predominant one that could be repeatedly isolated from both parental ambers. Hp17-5, which carries the 1580 base junctional fragment, was isolated only once. In addition, Hp17-5 is anomalous, although the same degree of amplification in genes 17 and 18 was observed, for unknown reasons less gene 17 mRNA and protein were made. It is therefore possible that

D. G. Wu and L. W. Black

780

I

2

3

4

I

2

3

4

kb

14 (b) (xl 493 -79

0 -27

(a)

(ll)

(a,@-

h)

(b)

Figure 8. Identification of the novel junctional fragment in the Hpl7 mutants by TaqI restriction enzyme digestion, and hybridization with 32P-labeled pCBR2 (genes 16 to 20) probe. DPjA from wild-type T4 (lane I), 17amNG178-alt (lane 2), Hp17-5 (lane 3) and Hp17-47 (lane 4) were digested with TagI restriction enzyme, electrophoresed on a I lh agarose gel and hybridized with 32P-labeled pCBR2 DNA. (a) Tap1 digestion profiles of T4 wild-type, gene 17 amber mutants and Hp17 mutants. (b) Hybridization patterns of (a) with 32P-labeled pCBR2 probe. The Tag1 restriction map of pCBR2 (genes 16 to 20) was constructed according to gene 26 t,o 19 sequence data (G. Mosig, personal communication; F. Arisaka, personal communication). Each restriction fragment on the autoradiograph has been assigned a, b, c, etc., corresponding to their locations on the map. Fragments x and y are probably derived from the gene 20 region, which has not been sequenced. kb, IO3 base-pairs. the H‘p17-5 mutant may have undergone changes in addition to the gene amplification. For example, the specific activity of the gp17 of Hp17-5 may have increased by an additional mutation, since significantly less protein (Fig. l(a), lane 2) functions as well as the amount in the Hp17-47-type mutant, as demonstrated by its plating efficiency and relative burst size on the ochre suppressor host (Table 1).

4. Discussion We have isolated numerous T4 gp17 hyperproducing mutants at high frequency through a simple one-step selection involving growth of gene 17 amber mutants in ochre suppressor strains of E. coli. The hyperproduction of gp17 observed in the infection by Hp17 mutants of an amber suppressor strain of E. coli was due to the

Ampl@cation

of Bacteriophe

T4 Genes 17 and 18

781

,I.50

0.93

(b)

(a)

Figure 9. Comparison of the junctional fragments in independently isolated Hp17 mutants by Southern blot analysis. The conditionsof Tug1restriction enzyme digestion, Southern blotting and hybridization with 32P-labeled pCBR2 probe were exactlv the sameas describedin the legendto Fig. 8. (a) TPdC DNA, 17amNG178-altand 4 initial Hp17 mutants. on CA165 su+B, while 9 to 15 were (b) 15 additional Hp17 isolates. Mutants 1 to 8 were isolated from 17amNG178-alt from 17amA465-alt. The missingbands at the lower part of the autoradiograph (b) (nos 4 to 9) were due to a poor Sout#hern transfer

of this portion

of the gel in this experiment.

amplification of DNA in the gene 17 and the adjacent gene 18 region. Two types of Hp17 mutants have been isolated, one predominating. Hp17-5, which appeared to be anomalous among all the Hp17 mutants, was distinguished from the others by the synthesis of less gp17, the co-hyperproduction of a 60,000 Mr partial polypeptide of gene 18, and the presence of a 1580 base TaqI junctional fragment. On the other hand, Hp17-47 exhibited the characteristics displayed by the majority of Hp17 mutants (28 of 29). In addition to an apparent hyperproduction of both gp17 and gp18 (Fig. l), a very similar, and generally nearly identical, Tag1 junctional fragment was found in these mutants (Fig. 9). We concluded that Hp17-47 carried an approximately 4300 base amplified unit containing the intact genes 17 and 18, while Hp17-5 carried a 3800 base amplified unit of gene 17 and a truncated gene 18 with the coexistence of the normal gene 18 in the same DNA molecule. The loss of gene 18 sequencesin Hp17-5 is likely a result of the fusion of two amplified units within the gene 18, probably at the C terminus. Southern analysis revealed that the Hp17 mutants,

kb, lo3 base-pairs.

although repeatedly purified from single phage plaques, actually constitute a heterogeneous population of phages with respect to the copy number of the amplified gene 17 and 18 sequences. The copy number varies from one to more than six copies. A single junctional fragment detected in both Hp17-5 and -47 also demonstrated that the amplified sequenceswere arranged as direct tandem repeats. Under a “head-full” DNA packaging mechanism, T4 DNA contains a permuted terminal redundancy of about 3000 to 4000 bases (1.8 +0*7 o/o of its 166,000 base genome; Kim & Davidson. 1974). This terminal redundancy was thought to be required for DNA replication (Mosig & Werner, 1969). A duplication of 3800 or 4300 bases as observed in the Hp17 mutants would almost or entirely eliminate terminally redundant sequences; therefore, phages carrying more than two copies of amplified sequencesare expected to be unable to propagate in the host unless a compensatory deletion is introduced into the phage genome, as described for T4 rI1 gene duplication (Homyk & Weil, 1974; Weil & Terzaghi, 1970). However, we detect no evidence

782

D. G. Wu and L. W. Hack

of such a deletion by restriction enzyme digestion and comparison of parental phage and Hp17 mutant DNAs, although this possibility is not absolutely excluded. It is probable, therefore, that such a duplication as an initial event of gene amplification in the Hp17 mutants (see below) does not completely abolish the terminal redundancy, and a short repeat at the end may remain and be sufficient. Or an unknown pre-existing deletion in the genome of the parental phage may allow the existence of one extra copy of genes 17 and 18, if the length of the amplification unit exceeds that of the terminal redundancy. A number of mechanisms have been proposed to account for the process of gene amplification in both prokaryotes and eukaryotes. In spite of a lack of direct evidence, “saltatory replication” has been one of the more popular models advanced for eukaryotes (Bullock & Botchan, 1981; Roberts et aE., 1983). In fact, multiple rounds of replication around origins as described in the saltatory replication model have been reported for T4 (Kozinski et al., 1980). However, the proximity of a replication origin is apparently required for such a mechanism, and no origin has, at least so far, been located near gene 17. Therefore, the involvement of the T4 mechanism of replication initiation in gene amplification of Hp17 mutants remains t’o be investigated. In prokaryotes, particularly in bact’eria, gene amplification is generally thought to result from an initial tandem duplication event. Both short and long repetitive sequences flanking the gene to be duplicated have been proposed to account for the frequent tandem duplications in E. coli and Salmonella by a RecA-dependent mechanism (Hill et al., 1977; Lin et aZ., 1984; Anderson & Roth, 1981; Edlund & Normark, 1981). In bacteriophage, gene duplication has generally been regarded as a consequence of illegitimate recombination. The frequency of lambda duplication was found to be independent of both phage and host recombination genes such as int, red and recA (Emmons et al.. 1975) and the endpoints of duplicated T4 rI1 and lambda genes were different among individual isolates (Parma et al., 1977: Emmons & Thomas, 1975). It is probable that gene amplification in the Hp17 mutants was also derived from an initial tandem duplication, which then leads to further amplification by normal homologous recombination. In a “gain and loss” exchange between two DNA molecules carrying tandem duplication, sequences of higher copy and single copy number are generated. The heterogeneity in the copy number of amplified sequences carried by the Hp17 mutants may reflect the result of such a process. An equilibrium population of amplified sequences (Fig. 6) presumably is produced by two countervailing forces. Molecules with increasing gene 17-1X copy number should act as non-viable “helper phages” to increase the phage yield in each burst through increased gp17 synthesis (see Table 1. m.o.i. 5 versus 0.1). At the same time, only phages

with two copies arr both viable and capable of giving rise to a second-round burst under the selection condition. In fact, if t,here is no mRNA terminat,ion between the fused genes IX and 17. then phages with higher 17-18 copy number may contribute disproportionately to tht, t,ot.al Hp 17 mutant output in each burst’. The observed difference between at least a fivefold increase in t,he level of gene 17 mRNA (Fig. 4) and a two- lo threefold overall increase in its gene dosage (Vie. 5) is readily explained if amplification brings the increased gene 27 sequences under ceontr‘ol of a stronger promoter of genes 18 and/or 19, while the weak gene 77 promoter does little t,o increase gene 18 or 19 mRNA synt,hesis. Indeed, this juxt,a,position may explain, at least in part). t.hc VOselection of gene 17 and 18 sequences as ;lr) amplified unit in such a selection, and t)hc> possibility of amplification without a compensatory deletion that we have apparently observed. It should be noted that the longer, and therefore t,hr more difficult to package. 4300 base amplification unit, which includes all of t’he gene IS sequenecs, is reiterated in preference to shorter unibs (e.g. Hp17-5). Compared with the other phage duplications. our results suggest that the initial event leading to the formation of the amplified sequences ill the HplT mutants may be significantly different. The 7’4 rl 1 duplication mutants were obtained at i lO\\ frequency ( 10V8) by a very strong selection (aross between t’wo different rI1 mutants carrying mutations overlapped that prohibited the formation of the wild-type recombinant (Weil e:t rcl., 1965). Hp17 mutants were isolat,ed from singlepoint mutations at an approximately lOO-fold higher frequency. A similar or identical junctional fragment’ observed in 28 out of 29 different isolates also indicated that the endpoints of the amplified sequences were not randomly located, as report,ed for T4 rI1, lambda duplication mutants, and the lac gene amplification mutants of E. toll: (Parma et CL!.. 1977; Emmons & Thomas, 1975: Tlsty Pt al.. 1984). Moreover. we have been unable to isolatti analogously similar Hp mutants in t,he adjacent genes 16 and 1X. These findings suggest that’ the mechanism could involve specific sequences. Since the T4 gene l&l9 region has been sequenced (G. Mosig. personal communication. F. Arisaka. personal communicat,ion). it, allowed us to sequenct’ the junctional fragment, of t#he common Hpl7-47 as well as the Hp17-5 mutants. By cxomparison with the known sequences, it, was drt,ermined that an imperfect homology of approximately 20 base-pairs located at t,he (I-terminal end of gene 76 and thfx beginning of gene 19 or within gene IX. respect,ively. are apparentSly involved in the amplification of genes 17 and IX in Hp17 mutants (unpublished are determining the junctional results). \Vr sequences of other independent, isolat,es of thtt H pl7 mutant, and also investigating the roles in this amplification event’ of specific T4 and E. co/i replication-reoombination fun&ions.

Ampli$cation

of Bacteriophage

This research was supported by National Institutes of Health grant AI11676 to L.W.B. We thank Drs A.-L. Chang, V. B. Rao and A. Zachary for their valuable suggestions and critical reading of the manuscript. We bhank Drs Mosig and Arisaka for providing the sequence data of the gene l&l!? region before publication.

References Albertini, A. M.? Hofer, M., Calos. M., Tlsty, T. D. & Miller, J. (1982). Cold Spring Harbor Symp. f&ant. Biol. 47, 841-850. Alt. F. W., Kellems, R. E., Bertino, *J. R. & Schimke, R. T. (1978). J. Biol. Chem. 253 1357-1370. Anderson, P. R. & Roth, J. R. (1977). Annu. Rev. M icrobiol. 3 1, 475-503. Anderson, P. R. & Roth, ,J. R. (1981). Proc. Nat. Acad. hi., Z:.S.A. 78. 3113-3117. Bellett, A. .J. D.. Busse, H. G. & Baldwin, R. L. (1971). In The Bacteriophage Lambda (Hershey, A. D., ed.), pp. 501-513, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Birnboim, H. C. & Daly, J. (1979). Nucl. Acids Res. 7, 1513-1523. Black, L. W. (1981). Virology, 113, 336-344. Black, I,. W. & Ahmad-Zadeh, C. (1971). J. Mol. Biol. 57, 71-92.

Brenner. S. & Beckwith, J. R. (1965). J. Mol. Biol. 13, 629-637. Bullock. P. & Botchan, M. (1981). In Gene AmpZi$cation (Schimke. R. T., ed.), pp. 215-224, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Cams, M. (1978). Nature (London), 274, 762-765. Chattoraj, D. K. &, Inman, R. B. (1974). Proc. Nat. Acad. Sci.. ljT.S.A. 71. 311-314. Edlund, T. & Normark, S. (1981). Nature (London), 292, 269-27

1.

Emmons, S. W. & Thomas, .J. 0. (1975). J. Mol. BioZ. 91, 147-l 52. Emmons, S. W., MacCosham, V. & Baldwin. R. L. (1975). ,J. Mol. Biol. 91, 133-146. GOB’. (‘. G. (1979). J. Viral. 29, 1232-1234. Hill. C. W.. Grafstrom, R. H., Harnish, B. W. & Hillman, B. 8. (1977). J. Mol. Biol. 116, 407-428. Homyk, T.. tJr & Weil, J. (1974). ViroZogy, 61, 503k525. Horvrtz. H. R. (1974). ,1. Mol. Biol. 90, 739-750. Howkes, R.. Niday. E. & Gordon, J. (1982). Anal. Biochem. 119, 142-147. Hsiao, C. I,. & Black, L. W. (1977). Proc. Nat. Acad. Sci., I:.S.A.

74, 3652-3656.

Just,

L., Frankis. R., Lowery, W. S., Meyer, R. A. & Paddock, (:. V. (1983). BioTechniques, Sept/Oct, 136-140. Kaufman, R. J., Brown, P. C. & Schimke, R. T. (1979). Proc.

Nat.

Acad.

Kim, J.-S. & Davidson,

Sk.,

U.S.A.

76, 5669-5673. 59. 93-I

N. (1974). Virology,

II.

T4 Genes 17 and 18

783

Kozinski, A. W., Ling, S.-K., Hutchinson, N,, Halpern, M. E. & Mattson, T. (1980). Proc. Nat. ilcad. Sci., U.S.A.

77, 5064-5068.

Kutter. E. & Rueger, W. (1983). In Bacteriophage T4 (Mathews, C. K., Kutter, E. M., Mosig. G. & Berget, P. B., eds), pp. 277-290, ASM Publications, Washington, DC. Kutter, E. & Rueger, W. (1985). 1985 Evergreen International T4 Meeting. Abstract. Laemmli, U. K. (1970). Nature (London), 227. 680-685. Laemmli, U. K. & Favre, M. (1973). J. Mol. Biol. 80, 575599.

Lin, R.-J., Capage. M. BE Hill, C. W. (1984). J. Mol. BioZ. 177, l--18. Long, E. 0. & Dawid. I. B. (1980). Annu. Rev. Biochem. 49, 727-764. Mattson, T.. Van Houwe, G., Belle, A., Seizer: G. & Epstein. R. (1977). Mol. Gen. Genet. 154. 319-326. Meyer, B. J., Kleid, D. G. & Ptashne. M. (1975). Proc. Nat.

Mosig.

Acad.

I%.,

U.S.A.

G. & Werner.

Z1.S.A.

64,

72, 4785-4789.

R. (1969).

Proe. Nat.

.4cad. Sci.,

747-754.

Parma, D. H., Heath, G. T., Che, C. C. & Annest, J. L. (1977). Genetics, 87, 593-619. Rigby, P. W. J., Dieckmann, M., Rhodes, C. & Berg, P. (1977). J. Mol. Biol. 113, 237-251. Roberts, J. M., Buck, L. B. & Axel. R. (1983). Cell, 33, 53-63. Schimke, R. T. (1984). Cell, 37, 705-713. Selzer. G., Belle, A., Krisch, H. & Epstein, R. (1978). Mol. Gen. Genet. 159, 301-309. Snustad. D. P. (1970). Virology, 41, 52-65. Southern, E. M. (1975). J. Mol. Biol. 98, 503-517. Spradling, A. C. & Mahowald, A. P. (1980). Proc. Nat. Acud. Sci., Z;.S.A. 77, 1096-I 100. Stark, G. R. & Wahl, G. M. (1984). Annu. Rev. Biochem. 53, 447-491. Synder, L., Gold, L. & Kutter, E. (1976). Proc. Nat. Acud. Sci., U.S.A. 73, 3098-3102. Tartof, K. D. (1975). Annu. Rev. Genet. 9, 355-385. Tlsty, T. D.. Albertini, A. M. & Miller, J. H. (1984). Cell, 37. 217-224. Towbin, H., Stahelin, T. t Gordon, J. (1979). Proc. Nat. Acad. Sci., Z!.S.A. 76, 4350-4354. Vanderslice, R. W. & Yegian, C. D. (1974). Virology, 60, 265-275. Weil. J. & Terzaghi, B. (1970). Virology, 42, 2:14-237. Weil, J.. Terzaghi, B. & Crasemann, J. (1965). Genetics, 52, 683-693. Wunderli, H.. van den Broek. J. & Kellenberger, E. (1977). J. Supramol. Struct. 7, 135-161. Young, E. T. & Van Houwe. G. (1970). J. Mol. Biol. 51, 6055619. Young. E. T., Mattson, T., Selzer, G., Van Houwe, G., Belle, A. & Epstein, R. (1980). J. Mol. BioZ. 138, 423-445.

Edited by M. E. Gottesman