., xxndu‘ad f. v. . ;‘4~\.-~\..‘ \;:._.- 4”: -‘v-‘»>\.‘.‘x‘.\. , . —~. ‘ 1' u >\-‘§ :Vl’\:.:~’\,» n.1,». . \x .31, wi‘\hh ‘3. iv“? V‘V IIHJ MI. 1:.)131. .4”. ,_ flung“. 4. ,4 l 'r 4:; 4 r r .4 'r, x I U 1 r V I (;(I'i 1 v I ‘u-i‘,‘ ,gx-I’a V 4 4 ..."" ‘:‘&~:\‘~_A. ' ,4“ ~ “_ ‘42::3‘ 441-: Q.“ “R", _.... -a, .5. . »y \ ‘w: .,. ,. 3.. 2w:- ' ’14" i ’ 41;, I~J , 47. ‘1 i 1 r l V‘ . ,e 4334': . ’4 K.) ‘ 5.; L.V\‘:.$§\h’f V‘S w 71%”. "’0 .§ :ff‘ “fir c 1 I 1 I 'v 1 F14 s. ‘_.. ~ ’. 4«J«.»-',.,\;, ‘- H»\V.\,"«\ ‘ ., 0:! ‘z. -\-V.-ux~:~~u\u ,. .~--.. .4 .4 :‘V ~~y54 * J" l .‘ L‘avgw ~ ’ "h" I n' '2‘...- ‘ ; .;4::-.' m: ‘ * ‘ ‘” V V“ 4 1' f;':’"‘;“’ $1.? mo. w. W ' \ .» , we .. ¢ b. “V“ V“ Q,- .. n. ...4 ”W J"? ”ENE”-.. "‘34:“ .nukwu'unxiv’ {r 0‘3?» {1:} VWgs h~~ ‘95:;55" 1:“? 3:55, ."lu‘\?“-‘« “‘3. —.‘ vi - 7 '“~ . :' 3:! -‘ . “71.2“."3'k‘ 3‘44 ,1}: #7:“? v it": “Xx-‘3." ‘1’"? V '2» .su-w-L . 1n- '- '- 3: »-|»v\§\ \xyc’u s3 ‘4’- 3: Wm “vi 4. -'.u ‘44-'32 ‘33..» '4’ ~- ' v- “: v\\ v 5“, . Nu.- hip-4 hvvuf 71.2“, "(~ku 7' " V‘ 4.4:: s .4 “A?" ‘ - y. _M,4"-.,\. ”k“ .1» ”4:?qu 4 \? {gawk 3'" 1' “M "“V‘ Mun-.nx‘n" .fik‘m‘fi?’ "HI-"w ~- \ ‘uyu§h-.‘ .. ‘ DKVIV‘E" . , 4442+’4“4“~:44:~v~ V - IVVI "A NH . 4,, “,a ”M”. "V” \ “ ,.~~y~.-7“.-A.. J _ "h m)“ 4 , \> ;x 44 -. "" ' """ ,V: V“ ‘f.§¢3:;- .12».\1T].~y 5,7,. I“: 7'!" "x‘hyya 3437’“. t . > n- my: ' , .4 ,. z , v.4 *‘4 ' wim‘?" \u. ,. s, V Fm," “‘flw‘ .I‘u ‘ ' -r I v‘ ’ w-g‘ $433. .2 ’ika If“. .. “fl“, swap-4 .\ Y' " ‘ KW ,.. y,\ y~nqq~rv ”\n ,4 QV V J“, ..\:‘.‘)‘s~34\\1.~1~§ n" 3’" ,- > {£4.11 . , ~34")? 3%?de 3*“ 45:1? .4“: In}? :53?" 7' .114}?! ‘ "3"”.‘1344 v 4 VI”). ‘5', .31)” H ' >33?) "5 W . _ \ y . .m J— - u: _ - r. ,a ’ ’ 4 x .4 , t . I 1 1 v M :5: '1 itfi-V'w \ 7 KIWI? 7V 439 ’5’ y." P‘p sake? V 31-415 \ . Y 4, '1‘: 1;: 4“ I“: ivy, 4 ‘ I i I . t I s 1‘ 4 ‘4 . fi', 'r'v‘v’ r‘_’fV)-;VQ’,F.. r7"- 1“ fP‘ Fatality: .1; n 4:: 4454454 EEK” "€44? m, 24-1-31? . '“.~- WW4 a4? «SM-34;: ~ A - ..-.' ’4'. » JMHrgnx‘ w:- " avg: v vq} dEva'- 4.3424444. .- ‘ i.” . 4- - . ' u; y?‘»a7 4,. \‘i iii , 4;?" m,» ‘44:, ”5??! i ’ 4/ l 1 r 5J4-~v J' 5;: '1 4 as $3 ’4; 1, 9' . WV. . . ' , s :1- .. 4 , J ”Pr >3hy‘5 $1; {9‘73}; . 4.”, '{Q’g‘r J' J 3 :4" fr”. 1'16“. .‘_, ..,.. ”‘5’?“"4CW i; ’35,]; ‘ -,--_,4.:V_:~€ ,. s: r ,4 >4, ,; .,,. 44mg, . .-L S? :14’4’4 r 4w": . 1211;421:1454 _.V..f,.:.:,;...4“; {’4 r{ 4f- :1"? r 3h "’ . 4 WM”? 2‘. 3?}. “w ,4 -,’- ° '1'. Jr” _ ALE/,3; (1”,, if,- 5’55). ”(4 ’ 7., .. 3);:h‘ '53-'11”! “:5 4' 5-5-3 44-! "f"’"’ 1?. "- 4.}, ,uigé “ r, '9 - 5:.“ ”nub,”- “£374 444’ 14.34;, ”J1”, a‘r13V'W3f my "2;! 3' -. "17:" ,' '71,}; fi-‘r’wr, v, )4 “5924:”, 7‘ r; .,‘,. x r Iv ’IM' '44.”: :rag’lf‘if’ ”3 .4! f' 4 "7‘7'3- ,{r‘st I P,» .. i t '4‘ 5916.4}; ",’ {1}” ,J’, “.4‘ fr». W” .1317. ,‘, 4" rrrhr'p‘z‘”, ‘ y--..¢‘;4 a". """' ‘ Jr . 'f‘Iv-W‘ '.» ‘1'L'f'1wa"! a‘flf. J I,” 1, ”jinx" “51:24." - '4 '4, I ’34:!" :{d’fi’fi‘Zifi- - 2:34:5": «Uh... " gr): {1" J". ' k 4 3".” " 3: I51" I "'4' ' ' ‘ r "’l' '13.} r7333: :r #1121 a” in? 4‘vPIJ'f ‘fw‘,',-'I4J 5.}"1; ”'44-‘11"43444‘2’r r14," - 1', #2," “jar"; J; . -.1 . .,.. ”5., 'r , fry-4 r_;~ p._m .. H; -.V‘:,.’ .532 5'" , ',- p 4" PH ‘1! ’J‘. {Han-1P5”, 5" 723V: 1‘ ‘-"..,J' .1, ‘l 4:! Vft]! ”in; N‘- ‘J’ _. 1: 4’1an ’ >'.Jo-" " 4 r J‘ 25"": :4: ’ "’ ' 4 r) 1:); :. gig: I 3"} 1,43": 7/»..413’: _:-”4-~,4 44"."! ”I J: I J” u A 44":i V ‘35:; ? T 7 ’Vr’l‘ ”#3 #114,344 ' 4 - "4r' ,. , J; r , . r“ 'P’;~’,!‘~',f‘ ‘7 Cliff-’1’? ,, J ‘4‘: ,. f 114,55" ” r -. 4'. I .. _ _ 44,4,» / 4:”1’52’444244224 45,4“ 4,:“9’24' , .1244 7’4 ”41/454’414’434'40’4'2’4‘" I4, 222,4 ,44 4,; J ‘4, 45:24: p; - ' “,4; J‘ .4:th ' , 4- “:4- 4 J fif'f .4’ (iii: J’h;,r,‘w’"l"“‘/.’n"j y’Jr: flaw ){u 4 ”7‘ _ r221“. Ira , ,.’ ,r!'l‘ m,‘ 'I‘,’ "1 . "JFFJ’H‘JI " Jr}? I" ‘rr/r V!» 7‘ rvr-r "‘ " I; (“'83 [7)" ."44’07; r4451? «:5; ' 4:"(1 '1}??? fifr}, 4 ’7': " "'4' ’ J {1,5451 Kn ' ’ I" ‘ f/f” 4’14". 31')?” ' ”4:545; ’ ' 7163’“, 7L; ' 1,}??er ‘,," -v,—r’r’r"-'.'l.f/f4/f I: "4.4-: {.4}? '4? 1.4 r . ' u l 4"_ 1 “1‘! {furrr r- ,4’4;44.-',‘rJ‘5;=’.;',.:-' $494474? 7,53, ”4:44 4'3? "5:2“; 4,”. K ’ 1"" ~/ in '.’ '- ' «4 4"!4 ’ ",J. 4/ 77444444495“?45,174,!) a!” ; ;§:""‘-'§ f, 4. Jr ,2, ":?'r/"'l’f"i {NJ/r; r r4 ,. , ,4’“ r . - “N. n! ‘ 4 -‘ - ’ nr :- t " ~ ’27:: 4;; J ,3” :44”1”'“'4 ’ I '9‘: I » , 4, 4r~'4' r51” '3?! ’ flak/44.1.44" lJ44 tr; )1” I’lrrij’“ [65?" '.”2’,€:"'f"",fg4 ”‘41:: [V‘s/n” rr'rnrvfl u!" “'11-” I'mn‘rrvr/ , ' .1, .t 4 rr .4 .4. 4 4".44 I” 4 ’ 4 '19 w?':.r,445'4”"'”'r4’>" " 6.53", H ’3’ I?” dig}; r' 4’ I” i" If I /rll' ’l/J. 12,.4/1 '4/"'/ EM!" I‘ 6"": HI} . .21 " ?’{&' ’IJ 41;; If?” V VV J’ :4. J . n.1,.” . 1.1.1” ,wr fr' up," 4"”: =4 . rm: .JJ-‘v’ Mfr ' {5,5,}? 4’21"? .f'n'” I, "LT ’ 4"?1‘" -"V {!,/,,4,";j4£:3k” :45! ”If 3"," 27:" p}. . ,4 an, ’97,. 71.1”” :53}: ”2'?ng 3"!" {I ‘ r' > "(h h ' '1! ,' . r 1 , . “f ”"53“",14‘44” '7" J5" ,. Z1444? M}, £4 '4. ..4,4.4’,4:::S’. 4’ ”I I I, J” In 4,44, I“; I H": IV); 9’} #21544! $7975! §r4§5§74i/’r;y‘ F'V'éj' 7;"; r} 41","; rélfffr p'": 1",». "”4"; 4:4, ,‘JJ . V 4 4,4,4], mm; 4,4,} 4,7,11,442'1', “1-,!" } ,JéJérVHJVVJy 9’15, 4,4: 41,147”; 451:;4'; 4’4, 1'1ny .- W'r : " ' . ",4 . ’4' AM ,, 5:744:34; f':4""">4”f";4>4'44414 . 4,, 4H,, ,“'-’444444,4,:¢~£,5,4 .,,,.,4,444/44~,M:4424;4:4:42’4>:4f4”4'44'4 ”074 ”,4 ,4, ,m', fir. {#074141 f ' ' ., a, 4;,,‘ [J 3,-4.4 74’4fo 4. ”fit? Jr, ”94’” , - , 44%”! (if if” 54:,” ”if”? 1.4/4”? low 44 /’4..4.u-.’f 1’ tr.‘ N .1” ' ,4, "/0, , V4144 :fl’a’l flair 'r'lr Ill, [1,. ;,JJ,,_ ,IJIH'I/ .l ‘ at” L/ M" I m I], mm W lIHIHHlflUfllD M “f 0 7" " 3|00635 419 LIBRARY Michigan State University This is to certify that the dissertation entitled MOLECULAR CHARACTERIZATION OF THE GOL REGION IN THE MAJOR CAPSID PROTEIN GENE OF BACTERIOPHAGE T4 presented by Kristin Jeanne Bergsland has been accepted towards fulfillment of the requirements for Ph , 1), degree in Microbiology A /7 = [X g g 7//« V / Major profes or Date /{/v/¢//Q?’ LOren R. Snyder MS U is an Affirmative Action/Equal Opportunity Institution 0-12771 MSU LIBRARIES m. w RETURNING MATERIALS: Place in book drop to remove this checkout from your record. FINES will be charged if book is returned after the date stamped below. ,W?’ ”7’ 2 ii 8 “ MOLECULAR CHARACTERIZATION OF THE GOL REGION IN THE MAJOR CAPSID PROTEIN GENE OF BACTERIOPHAGE T4 BY Kristin Jeanne Bergsland A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Microbiology and Public Health 1987 ABSTRACT MOLECULAR CHARACTERIZATION OF THE GOL REGION IN THE MAJOR CAPSID PROTEIN GENE OF BACTERIOPHAGE T4 BY Kristin Jeanne Bergsland Overproduction of the Escherichia coli lit protein, encoded by the cryptic DNA element e14, can block gene expression late in infection by bacteriophage T4. This defect can be totally overcome by T4 ggl mutations, which map within a 40 bp region near the amino-terminal end of gene 23, the major capsid protein gene of T4. The defect in gene expression is caused by the interaction of the lit protein with part of the T4 major head protein, a short amino acid sequence of no more than 25 amino acids from the ggl region. The expression of translational fusions of the wild—type ggl region with lag; is severely inhibited by the presense of lit_protein. Therefore the interaction causes a termination of translation (or transcription) of gene 23 or other genes into which the g9; region is cloned. Replication of a plasmid containing the g9; region can also be inhibited by the lit protein. In addition to these gig-effects on gene expression and plasmid replication, the interaction of lit with the g9; region can block the expression of other genes in trans. A clone of the wild-type gol region in a lambda vector causes the growth of lit—overproducing bacteria to rapidly decrease and eventually cease, while there is no such effect on a strain without the lit gene. g9; mutations are single base pair changes which relieve both the gig and trans effects of interaction with lit. Nucleotide sequence analysis of thirteen go; mutations suggests that not all mutations change the amino acid sequence. I speculate that some may function by altering an RNA secondary structure in the go; region, which is required for termination. The interaction of the lit protein with the go; region of T4 may have exposed a normal regulatory mechanism whereby the synthesis of the late gene products is coordinated with assembly into T4 heads. This may involve a novel type of retroregulation, in which a nascent polypeptide can block the transcription (or translation) of a gene and also trigger the inhibition of other gene expression in trans. To Mom and Dad, whose support and encouragement made this possible. iv Acknowledgements I wish to acknowledge the advice and guidance given by the members of my committee: Drs. Michele Fluck, P.T. Magee, Leonard Robbins, and Barry Chelm. I especially want to thank Dr. Larry Snyder for his patience, for teaching me to think as a scientist, and also for his unending optimism and enthusiasm which made this work much easier. I wish to thank all of my Giltner Hall friends, especially the Breakfast Club, for scientific and moral support over the years and for making this such a pleasant experience. Finally, I wish to thank Tim Caffrey, who was with me from the start and was a great source of strength and encouragement. Table of Contents List of Tables List of Figures Introduction Literature Survey T4 and the History of Molecular Biology The T4 Genome Prereplicative Transcription Early Middle Anti-late Postreplicative (late) Transcription RNA Polymerase Modification Coupling of Transcription and DNA Replication The Effect of Nucleotide Modification on Late Transcription In-vitro Systems Expression of Cloned Late Genes In Vivo Post—Transcriptional Regulation Translational Regulation RNA Splicing T4 Head morphogenesis Manuscript Summary Appendix Bibliography vi 12 13 14 15 17 18 19 20 22 22 24 25 29 86 91 95 List of Tables Table Page 1 Bacterial Strains 33 2 Plasmids 51a vii Figure la List of Figures Genetic and restriction map of the gene 23 region of bacteriophage T4. T4 late protein synthesis after infection of 1it(Con) E. coli by gene 23 amber mutants. DNA and amino acid sequences of gene 23 in the gol region showing mutational changes and deletion endpoints. Schema of structure of translational fusions with T4 gol sequences fused to lacZ. Expression of gol-lacZ fusions and plasmid replication in 1it(Con) bacteria at 30°C. Plasmid replication and expression of gol- lacZ fusions in 1it(Con) bacteria first grown at 42°C, then shifted to 30°C. A possible secondary structure in the gol region of gene 23. Growth of E. coli infected with a lambda clone of the T4 gol region. viii SO 55 61 7O 74 80 94 INTRODUCTION Despite the wealth of knowledge about regulation of gene expression in bacteriophage T4, many important questions concerning the control of expression of the T4 late genes remain unanswered. The late genes are thought to be regulated primarily at the level of transcription, and expression of these genes is coupled to concurrent DNA replication. The products of the late genes are largely the virion structural components, which are rapidly and precisely assembled into 100-200 new phage within 25—30 min. after infection. Assembly of the viral capsid is a complex process, requiring that exact intracellular ratios of the components be maintained to achieve proper capsid structure. It is not known how the balanced synthesis of the late gene products is accomplished, but it seems likely that mechanisms exist which coordinate synthesis of the protein components with assembly into phage structures. In general, T4 utilizes the same types of regulatory processes that control the more complex genomes of its E. ggEE host and other prokaryotes. In addition to the transcriptional mechanisms that control phage development, some T4 genes are regulated at the levels of translation, antitermination and RNA processing. Indeed, the first example of RNA splicing in a prokaryotic system was the thymidylate synthase gene of T4. Therefore, elucidating 2 regulatory mechanisms which operate in T4 phage may add to our understanding of the mechanisms which control gene expression in other prokaryotes and, perhaps, in higher organisms as well. One approach that our laboratory has taken to the study of T4 late gene expression, has been to analyze strains of E. ggli in which late gene expression is defective. One such strain is E. ggl; CTer, which blocks DNA replication and late gene expression of polynucleotide kinase- and RNA ligase-deficient mutants of T4 (Sirotkin et al., 1978; Jabbar and Snyder, 1984). A search for a similar restricting strain of E. 99;; K-lZ resulted in the identification of ii: mutants (Cooley et al., 1979). E. ggEE El: mutants restrict the growth of T4 polynucleotide kinase mutants at 37°C, but in addition, even the growth of wild-type T4 is restricted at temperatures below 34°. T4 fails to multiply on lg: mutant hosts due to a severe block in late gene expression (but not DNA replication) at the restrictive temperature. The 1;: gene maps at 25 min. on the E. ggli K-lZ genetic map and is located within the cryptic DNA element e14 (Kao and Snyder, submitted). e14 is a l4kb element which is found in most K strains of E. 99;; and can excise from the chromosome upon induction of SOS functions (Greener and Hill, 1980). The E1: gene is nonwessential for T4 development, since T4 multiplies normally on E; ggli strains which lack e14. The mutations that block T4 late gene expression are 3 called EEE(Con), and cause the overproduction of a 34 kD protein which is localized in the bacterial inner membrane (Kao et a1. 1987; Kao and Snyder, submitted). T4 mutants have been isolated which can form plaques on a EE§(Con) strain at the restrictive temperature. These T4 mutants are called ggE, for they grow gn EEE(Con). E9; mutations are located within a 40 bp region near the amino— terminal end of the major capsid protein gene of T4, gene 23 (Champness and Snyder, 1984). These mutations are gig-acting for expression of gene 23, and possibly other closely linked genes, at late times after T4 infection of a EE:(Con) host (Champness and Snyder, 1982). In addition to its effect on T4 growth, the g9; region cloned in a plasmid can affect plasmid transformation in a way which is analogous to its effect on T4 gene expression. A plasmid which contains the wild-type ggl sequence will not transform a EE:(Con) host. However, if the plasmid clone contains a g9; point mutation or is deleted for the g9; region, it can transform a EE:(Con) strain efficiently (Champness and Snyder, 1984). When the work presented in this dissertation was initiated, the objective was to characterize the T4 ggl region in an effort to understand the molecular basis of its activity. Among the important questions that we wanted to address were: How large is the g9; site and what DNA sequences are required for activity? Is transcription or 4 translation required for activity? What is the molecular nature of the changes caused by g9; mutations? How does the ggl site exert its effect on gene expression? At what level does the ggE site act to block transformation of plasmid clones into a 1;: (Con) strain? In this paper, evidence is presented that a short polypeptide sequence of about 25 amino acids within the gene 23 protein is responsible for the phenotypes. The lg: protein presumably interacts with this amino acid sequence which results in two distinct effects: a gig-inhibition of expression of gene 23 due to a termination of translation (or transcription) in the ggl region, and a Egggg inhibition of expression of other genes in E. ggll. The interaction can also cause a gig-inhibition of the replication of a plasmid in which the g9; site resides. The literature survey reviews the main types of regulatory mechanisms which are known to function in T4. Because the expression of the late gene products may be affected by their role in phage capsid assembly, a discussion of T4 head morphogenesis is presented. The experiments described in the appendix are not for publication, but support the data in the manuscript and the conclusion that the cloned region has a Egggg effect on the expression of other vector and host genes. Literature Survey T4 and the History of Molecular Biology Bacteriophage T4 is large and complex, with over 150 genes in its genome, yet it is one of the most well- characterized of all viruses. Since its discovery in the early 1940s, T4 has been the focus of extensive investigations and has played a vital role in the birth and growth of the science of molecular biology. Early studies of the viral life cycle provided a foundation for subsequent work on genetic structure and recombination. Delbruck and Bailey (1946) first reported that genetic information could be exchanged among different viruses which multiply within the same cell. Luria (1947) proposed a recombination-type mechanism to explain these results. Soon after, genetic linkage and crossing over were clearly demonstrated by Hershey and Rotman (1948). The role of DNA in encoding the genetic information of the virus was later shown by the experiments of Hershey and Chase (1952). The general concept of the nature of the gene was greatly influenced by the work of Benzer on the T4 rIIA and rIIB genes. His genetic fine—structure analysis and the extent to which he achieved mutational saturation of the two genes may never be surpassed. He developed the first conditional—lethal system to be used with T4, that is, the lethality of rII mutants on lambda lysogens in contrast to their excellent viability on strains without lambda. Benzer 6 (1959) used this system to demonstrate that the gene is a linear structure. The premier contribution of T4 to molecular biology must be its role in establishing the nature of the genetic code. Crick et a1. (1961) used a powerful genetic approach to the problem, taking advantage of the well-defined T4rII system developed by Benzer. In elegant experiments using frame shift mutations and suppressors in the rIIB gene, Crick and co-workers showed that the genetic code is read in triplets from a fixed starting point, that the code is non— overlapping and that it is degenerate. The colinearity of the gene with the polypeptide it encodes was subsequently proven (Sarabhai et al., 1964) by demonstrating that the order of amber mutations within a gene as predicted by the amber peptide fragments produced was identical to the order found by conventional 3—point genetic mapping. Another achievement of T4 is that the existence of mRNA was first demonstrated in T4-infected E. 991;. Brenner et a1. (1961) showed conclusively that the informational intermediate between DNA and protein could not be ribsomal RNA, and that the RNA which was newly synthesized after T4- infection became temporarily associated with ribosomes constructed before infection, as did the nascent T4 proteins. The first description of any restriction-modification system also resulted from work with T4 (Luria and Human, 1952). The phage DNA is normally glucosylated, and when 7 grown on a host mutant which is unable to supply the precursors for the glucosylation reaction, the progeny phage DNA becomes susceptible to restriction in other E. ggli strains (Hattmann and Fukasawa, 1963). This restriction of unglucosylated phage DNA is now known to be due to a specific restricting system called the Rgl system (Revel, 1967). Thus, modern molecular biology is solidly based on research with T4 and its host Escherichia coli. Much is understood about this remarkable organism, yet there are many intriguing questions which still remain to be answered. The T4 Genome The T4 genome is composed of 166 kilobase pairs of DNA. The mature DNA molecules in the phage particles are linear and about 170 kb in length, due to a 2.3% terminal redundancy. The DNA is packaged by a "headful" mechanism which gives rise to a population of phage carrying all possible circular permutations of the genome. Thus the genetic map is circular even though the virion DNA is linear (Streisinger et al., 1964). Approximately 150 of the T4 genes have been identified and characterized, which represents about 90% of the coding capacity of the genome (Mosig, 1983). The genes are grouped into two major classes according to when they are expressed during infection. Prereplicative genes are expressed early in infection and encode products which function in phage DNA 8 metabolism. Postreplicative genes are expressed later in infection, after the start of DNA replication, and encode phage assembly functions. The complex developmental program of bacteriophage T4 is regulated, for the most part, at the level of transcription. Three main classes of transcription units have been distinguished in T4 (early, middle, late). Each of these begins to be expressed at a different time after infection, is served by a distinct class of promoters and requires a different form of RNA polymerase. The early and middle modes of transcription are initiated in the prereplicative period. Early gene expression begins immediately after infection, while middle transcription starts about 1.7 minutes post—infection at 30°C. Late transcription begins 1- 2 minutes following the onset of DNA replication, which usually occurs at 5—6 minutes after infection. Transcription of the early genes is shut off about the time late protein synthesis begins, but some prereplicative genes continue to be expressed late in infection due to residual middle mode transcription (Rabussay, 1983). The T4 genes are arranged on the genome such that prereplicative transcription units are found mainly in two large blocks and the late transcription units are clustered in three major groups, separated by the early clusters (Wood and Revel, 1976). Prereplicative genes make up more than half the genome, and they are all transcribed with the same 9 polarity, off the "1" stand. The late genes are all transcribed in the opposite direction, off the "r" strand (Guha et al., 1971). Prereplicative Transcription Early Transcription Two classes of genes are found within early transcription units. The first is expressed immediately after T4 infection and is referred to as immediate early (IE). Transcription of the second group begins about 2 min. after infection and these genes are called delayed early (DE). The early transcription units are served by early promoters, which are recognized by unmodified E. 99;; RNA polymerase holoenzyme (Brody, Diggelmann, and Geiduschek, 1970). Early promoters contain strong homologies to the -10 and the —35 regions of E. 99;; and other early bacteriophage promoters (Rosenberg and Court, 1979). Early T4 transcription units somewhat resemble constitutive bacterial operons. The promoters are E. ggli— like, they encode polycistronic messages and end with rho— independent termination sequences. The IE genes are proximal to early promoters and DE genes are distal (Salser et al., 1970). IE and DE genes are separated by potential transcription termination sites. Protein synthesis is absolutely required to allow RNA polymerase to transcribe past IE—DE junctions. When T4 10 infects E. 99;; in the presence of chloramphenicol (CAM) early transcription units produce only short RNAs, which are promoter-proximal transcripts (IE RNA), and the DE genes are not transcribed (Brody et al., 1970). The CAM effect on early transcription is due to Egg-mediated polarity, analogous to that seen in bacterial operons. Lack of translation of the IE RNA allows 5E9 factor to recognize 5E9 binding sites on the RNA and induce termination. Mutations in the E. ggli 5E9 gene prevent this premature termination and allow the synthesis of normal amounts of DE RNA after T4 infection in the presence of CAM (Daegelen et al., 1982). T4 has developed a way to overcome this potential termination, since RNA chains are elongated from IE to DE regions in a normal (£Eg*) infection. Two mechanisms which would allow this anti-termination have been proposed. The first involves a T4 equivalent of the phage lambda N protein. T4 mutants have been isolated called 9999 a which are capable of growing on an E. ggii strain with an unusual 3E9 mutation, £229 (Caruso et al., 1979). This mutant Egg is insensitive to the normal T4 anti—termination mechanism and no DE transcripts are seen after wild-type T4 infection. The T4 ggggd mutation overcomes the effect of :gEE, but the function of ggggd as a true anti-terminator protein has yet to be experimentally proven. In the second model for anti-termination, transcription beyond potential rho sites requires only the act of ll translating the IE RNA, and not a particular N—like protein. Transcription and translation of the IE RNA are known to be closely coupled, and it is possible that the movement of ribosomes along the nascent IE RNA is all that is needed to prevent 5E9 from acting at the IE-DE junctions (Brody, Rabussay, and Hall, 1983). This model is supported by evidence that DE transcription is normal after T4 infection in the presence of amino acid analogs which yield inactive proteins. This suggests that transcription past potential 3E9 sites does not depend on the synthesis of functional IE proteins (Thermes and Brody, 1984). Whether or not a T4 protein is required for anti- termination, it is clear that T4 loses its susceptibility to CAM—induced polarity between 1.5 and 4 min. after infection (Brody, Sederoff, Bolle, and Epstein, 1970). Termination sites apparently change during the course of infection such that 3E9 cannot act at these sites even if protein synthesis is eliminated. The change of a 3E9 site from CAM-sensitive to CAM-insensitive is stable during many rounds of transcription and is irreversible. It has been proposed that a host or phage protein binds to the DNA after the transcription—translation complex moves past a 3E9 site, which serves to decrease subsequent RNA polymerase pausing and eliminate polarity. Alternatively, a protein could bind to RNA polymerase to alter its termination properties (Brody, Rabussay, and Hall, 1983). 12 Middle Transcription O'Farrell and Gold (1973) demonstrated that some T4 prereplicative genes were under the control of promoters not recognized until 1-2 min. after infection. RNA synthesis from middle transcription units is initiated at middle promoters and is under positive control by the product of the T4 £935 gene (Mattson et al., 1974). Early and middle transcription units overlap to a great extent, and many prereplicative genes are expressed in both an early mode and a middle mode. Middle promoters are often found between IE and DE genes. It has been demonstrated in infections of E. coli tabC hosts (blocked in early-mode DE expression) that almost all DE genes have a middle (mg:- dependent) mode of expression (Pulitzer et al., 1979). The $935 gene function is not essential for T4 growth on a wild-type E. 991; host, but in a PEPE strain, the expression of some genes is totally ngE-dependent. Pulitzer et a1. (1985) have recently identified two new T4 genes, mgtE and 3939, which control an alternative mode of middle transcription. The function of these gene products is unclear, but it is thought that one or both might play a role in anti—Egg antitermination between IE and DE genes, and they are unable to function in a :gEE host. In fact, ggtg may be equivalent to the ggmgg gene described above. The motA gene product is probably involved in middle 13 promoter recognition. The 3935 gene product has been shown to be an IE protein which binds to DNA (Uzan et al., 1983) and increases the affinity of RNA polymerase for T4 DNA (de Franciscis and Brody, 1982). Middle promoters show strong homology to the -10 region of E. ggli promoters, but there is little similarity in the -35 regions. There may be a new consensus sequence in the -35 region, though, consisting of the sequence AiTGCTT. This sequence may comprise part of the $935 recognition site (Brody, Rabussay, and Hall, 1983). Although it is fairly certain that 3935 helps facilitate RNA polymerase initiation at middle promoters, it has also been suggested that it may function in antitermination at IE—DE junctions. Elucidation of the roles of the mg: proteins awaits further experimentation. Anti—late Transcription Anti-late RNA is transcribed, like nearly all early and middle RNAs, from the "l" strand of DNA, but from a region of the chromosome where late genes are located. This RNA is therefore capable of hybridizing with the late RNA, giving rise to RNA-RNA duplexes (Geiduschek and Grau, 1970). Anti- late RNA accounts for 2-3% of total prereplicative RNA, and it is transcribed from widely dispersed regions of the chromosome (Young et al., 1980). Most evidence suggests that anti-late RNAs are transcribed in early and middle modes. However, some anti- 14 late transcripts have been identified which are coordinately regulated with late RNA (Elliott, Kassavetis, and Geiduschek, 1984). These transcripts map to the vicinity of genes 22, 23, and 24 (capsid protein genes) and the 5' ends are positioned to generate convergent transcription across the late promoters. The anti—late RNAs are much less abundant than the gene 22, 23, and 24 mRNAs. Anti—late RNA is found in polysomes (Helland, 1979) and a number of the transcripts contain open reading frames. To date, there are no examples of proteins encoded by anti-late RNAs, and the significance of these transcripts remains a mystery. Postreplicative (late) Transcription The post—replicative period begins at about five minutes after infection at 30°C, and late transcription starts one to two minutes later, after the onset of DNA replication (O'Farrell and Gold, 1973; Young et al., 1980). Many of the studies of late transcription have focused on genes which encode proteins involved in T4 head morphogenesis, since these genes are highly expressed at late times in infection. In fact, gene 23, the major capsid protein gene, codes for the most abundant protein made during the late phase of infection, constituting 20-30% of the late protein in cells (Young et al., 1981). Studies of the transcription pattern in the gene 23 region have revealed that gene 23 is transcribed from at 15 least 3 promoters (Kassavetis and Geiduschek, 1982). Approximately 80% of the gene 23 transcripts represent monocistronic messages, and the rest are located at the distal end of polycistronic RNAs which include sequences from the upstream genes 21 and 22 (Young et al., 1981). Analysis of the temporal appearance of these transcripts showed that each of these genes has its own promoter which is activated between 6 and 11 min. after infection (Elliott et al., 1984). The promoter sequences of several late genes, including genes 21, 22, 23, and 24, have been determined (Kassavetis and Geiduschek, 1982; Christensen and Young, 1982). They all share a common sequence, TATAAATA, which has been dubbed the "juke box" in the region of E. 99;; -10 sequence, but are quite dissimilar in the -35 region (Christensen and Young, 1983). Deletion analysis of the gene 23 promoter has shown that a -35 region is not required for recognition by RNA polymerase (Elliot and Geiduschek, 1984; Volker et al., 1984). Furthermore, Elliott and Geisduschek (1984) have demonstrated that a 35 bp DNA fragment from the region of the gene 23 promoter, from -18 to +17 with respect to the transcriptional start site and containing a "juke box," was fully competent to serve as a T4 late promoter. RNA Polymerase Modification Although there is only one class of late genes based on temporal patterns of late transcription, the requirements for 16 late transcription are much more complex than for.early gene expression. Late transcription depends on modification of RNA polymerase and a special competent (processed) DNA template usually created in the course of T4 DNA replication. The first change in RNA polymerase after T4 infection is ADP-ribosylation of the alpha subunits by the products of the T4 213 and Egg genes (Goff, 1974). Subsequently, five T4- specified proteins bind to RNA polymerase. These include the products of genes 33, 55, and 45 (Horvitz, 1973; Ratner, 1974) and two smaller polypeptides of 10K and 15K (Stevens, 1972). The gene for the 15K protein has recently been identified and is called £295 (Williams et al., 1987). The products of genes 33, 45, and 55 are known to be directly involved and continuously required for late transcription (Bolle et al., 1968; Coppo et al., 1975). These modifications are thought to be involved in turning off bacterial and/or T4 early gene transcription. By 9 min. after infection, host transcription is no longer detectable and T4 late transcripts become the dominant species. Most of the T4-induced changes antagonize the binding of the host sigma subunit to the core RNA polymerase, which results in decreased recognition of early promoters (Rabussay, 1983; Stevens, 1977; Malik and Goldfarb, 1984; Ratner, 1974b). It has been shown that the host sigma factor is not required for recognition of T4 late promoters (Kassavetis and Geiduschek, 1984). In fact, the product of 17 T4 gene 55 substitutes for sigma as a specificity factor for RNA polymerase recognition of late promoters (Malik et al., 1985). All of the T4 proteins which bind to RNA polymerase, except £225, are present in substoichiometric amounts. This implies that RNA polymerase complexes of different subunit composition can coexist in the infected cell. This would allow simultaneous transcription of different classes of genes late in infection. Thus, some early and middle mode transcription can persist late in infection although transcription from late promoters is dominant (Rabussay, 1983). Coupling of Transcription and DNA Replication In addition to modification of the RNA polymerase, there is a structural constraint on the DNA template for activation of late transcription. Late gene expression is normally coupled to concurrent DNA replication (Riva et al., 19703), since blocking replication invariably prevents late transcription. However, it is possible to uncouple these processes by introducing T4 mutations (DNA polymerase, DNA ligase and gene 46 exonuclease) which lead to formation or stabilization of breaks in the unreplicated DNA (Wu et al., 1975; Riva et al., 1970b). It has been proposed (Geiduschek et al., 1983) that the requirement for "competence" of the DNA template for late transcription results from the 18 inability of RNA polymerase to recognize late promoters in intact double-stranded DNA. Alteration of the DNA structure during replication, or alternatively, the introduction of nicks or gaps in the DNA makes late promoters accessible to RNA polymerase. The Effect of Nucleotide Modification on Late Transcription T4 DNA normally contains 5-hydroxymethylcytosine (HMC) in place of cytosine. In addition, the DNA is glucosylated by attachment of a glucose moiety to the hydroxymethyl group of HMC. This allows the phage to distinguish between its own and foreign DNA during the shutoff of host functions early in infection, and also allows it to avoid the Rgl restriction system (Revel, 1967) which is specific for HMC-DNA. T4 can be forced to incorporate cytosine instead of HMC into its DNA through a combination of phage and host mutations (Kutter et al., 1975; Runnels and Snyder, 1978). However, late gene expression from these phage with cytosine DNA (C-DNA) is severely impaired (Kutter et al., 1975; Wu and Geiduschek, 1975). A T4 gene product is responsible for blocking late transcription from C—DNA. Mutations in this gene, gig, allow late transcription from DNA with cytosine (Snyder et al., 1976). The T4 gig gene product is also responsible for unfolding the host nucleoid shortly after infection (Snustad et al., 1976; Sirotkin et al., 1977) and is involved in shutting off some host RNA synthesis (Sirotkin 19 et al., 1977). The mechanism of action of gig in blocking transcription remains obscure. There is evidence to suggest that gig can interact with both the DNA template and the RNA polymerase. Sirotkin et al. (1977) first suggested that gpgig is an RNA polymerase-associated protein. Pearson and Snyder (1980) later proposed that gig does not act simply as an RNA polymerase binding protein, but rather, acts on specific DNA structures to prevent utilization of promoters and cause premature termination of transcription. By this model, the gig protein normally unfolds the nucleoid, but cannot distinguish T4 C-DNA from E. ggii C-DNA, and thereby destroys T4 DNA structures required for late transcription. More recent experiments have shown that the gig protein does bind to DNA in the absence of RNA polymerase (Snustad et al., 1986). However, Snyder and Jorissen (manuscript submitted) have isolated E. ggii mutants which prevent gig function and propagate T4 with C-DNA normally. These mutations map to the B-subunit of RNA polymerase, providing genetic evidence for association of alc with RNA polymerase. In vitro Systems Late transcription has historically been difficult to analyze ig vitro due to the complex requirements for template and RNA polymerase modification. The first system which was reliable and highly selective for late transcription was 20 based on concentrated, gently prepared lysates of T4-infected cells supported on cellophane disks (Rabussay and Geiduschek, 1977a). This system was dependent on prior ig yiyg replication of the T4 DNA, and the competence for late transcription was very sensitive to mechanical disruption. More recently, Kassanetis et a1. (1983) have developed an ig giggg transcription system using plasmid clones of T4 late genes and RNA polymerase purified from E. ggii at late times after T4 infection. They have demonstrated efficient and specific initiation of transcription at several late promoters, in the absence of DNA replication. The plasmid DNA must be negatively supercoiled for optimal results, presumably to activate the template, which would normally be accomplished by DNA replication. Further experiments using this system have proven that the E. ggii sigma subunit is not required for RNA polymerase initiation at late promoters, and that the T4 gene 55 product alone suffices to confer late promoter specificity on unmodified RNA polymerase core enzyme from uninfected bacteria (Kassavetis and Geiduschek, 1984). Expression of Cloned Late Genes In Vivo The expression of cloned T4 late genes has also been studied ig yiyg after superinfection with T4 phage. Cloned late genes can complement mutations in the infecting phage (Mattson et al., 1977; Jacobs et al., 1981) providing that the phage contains other mutations to prevent degradation of 21 the plasmid DNA (T4 Eggg and gggE genes encode endonuclease functions). An gig mutation in the phage also aids complementation by allowing transcription from cytosine- containing DNA. Studies of the expression of cloned late genes have revealed differences in regulation between genes on a plasmid and in the phage chromosome. In studies of the cloned gene 23 (capsid protein gene), it was demonstrated that expression of the plasmid gene was dependent on T4 genes 33, 45, and 55, as are the chromosomal late genes, but the expression was substantially independent of DNA replication (Jacobs and Geiduschek, 1981). In addition, expression of the plasmid gene required the activity of gene 46 (DNA exonuclease), which is normally not required for T4 late transcription. "It has been proposed (Geiduschek et al., 1983) that the involvement of gene 46 with the expression of plasmid clones represents a novel mechanism of template activation. There is some evidence from maxicell experiments used to detect plasmid-encoded proteins that a plasmid clone of gene 23 is expressed in uninfected bacteria. However, subsequent studies have shown that the late promoter is not used in uninfected cells (Volker et al., 1984; Kassavetis et al., 1984) and that expression of the cloned gene in pBR322 is mediated by the g—lactamase promoter (Kassavetis et al., 1984). 22 Post-transcriptional Regulation Translational regulation In addition to the transcriptional mechanisms which control T4 development, some genes are also regulated at the levels of translation and RNA processing. Translational regulation will be considered first. T4 is thought to control translation in two ways: first, by a general modification of the translational apparatus which enhances the specificity for T4 mRNA and second, by specific regulation of individual genes with translational repressor proteins. The translation of non-T4 mRNA is severely inhibited after T4 infection. This is true both for other phage RNAs (Hattman and Hofschneider, 1968; Pearson and Snyder, 1980) and E. ggii mRNA (Kennel, 1970). Hsu and Weiss (1969) discovered that ribosomes from T4-infected cells were much less efficient for ig yiigg translation of phage M82 RNA and g. Eli RNA. Some evidence exists for chemical modification of the ribosomal proteins 81 (Rabussay and Geiduschek, 1977b) and $12 (Artman and Werthamer, 1974) after T4 infection. Since the ribosomal protein 81 interacts with the translational initiation factor IF3, the T4-modified 51 may interfere with IF3 initiation activity on non—T4 RNA (Wiberg and Karam, 1983). It has also been demonstrated that there are T4— specific proteins tightly bound to ribosomes isolated after 23 infection (Smith and Haselkorn, 1969; Dube and Rudland, 1970). However, the identity and function of these proteins have not been determined. The level of translation of individual T4 mRNAs may be determined largely by the strength and accessibility of the ribosome—binding site (Wiberg and Karam, 1983; Shinedling et al., 1987). For example, the T4 lysozyme gene is transcribed early in infection, but not translated due to formation of an RNA secondary structure which blocks the ribosome binding site (McPheeters et al., 1986; Knight et al., 1987). This prevents premature synthesis of the enzyme, which is made at later times in infection when the gene is transcribed from a different promoter. However, the use of some mRNAs is regulated by specific protein repressors of translation. One example is the gene 32 helix-destabilizing protein, which is known to be autogenously regulated at the level of translation (Gold et al., 1976). This protein binds to single stranded DNA during replication, and when this substrate is limiting it binds to its own mRNA to repress further translation. Another T4 translational repressor protein, Eggg, not only regulates its own synthesis (Cardillo et al., 1979) but plays a role in the utilization of a subpopulation of T4 early transcripts. It is estimated that 10-15% of T4 early proteins are normally under translational regulation by the regA gene product (Miller et al., 1987). Karam et al. 24 (1981) have shown that for the rIIB gene, the target site for mRNA binding of gggg overlaps the ribosome binding site. Thus, the gggg protein probably functions by competing with ribosomes for the same site on the mRNA. Interestingly, gggg can also apparently stimulate the synthesis of certain phage and host proteins, by an unknown mechanism (Wiberg and Karam, 1983; Miller et al., 1987). The significance of the global nature of gggg regulation is unclear, but it may be part of a coordinated mechanism for control of early development in T4 phage infection (Wiberg and Karam, 1983). RNA Splicing Prokaryotic mRNA splicing was first demonstrated with the thymidylate synthase (3g) gene of T4 (Chu et al., 1984). Splicing is an autocatalytic event which involves excision of a 1017 bp intron, which is then found in circular form (Chu et al., 1985). Mechanistically, the 3g intervening sequence is a member of the group I self-splicing introns, which have been identified in Tetrahymena and in the mitochondria of other eukaryotes (Chu et al., 1986). It has been observed recently that T4 contains multiple self-splicing introns (Gott et al., 1986). A new intron of this type has been discovered in the gigE gene, which encodes ribonucleotide diphosphate reductase (Gott et al., 1986; Sjoberg et al., 1986). There is evidence to suggest that introns may be found in several other T4 genes, but these 25 genes have yet to be identified. Both of the genes which are known to contain introns participate in the same metabolic pathway, i.e., synthesis of DNA precursors. It has been proposed that splicing may play a regulatory role in determining nucleotide pool sizes (Gott et al., 1986). It is worth noting that regulation at the level of translation and RNA processing have, to date, only been observed for T4 genes expressed during the prereplicative period. It is not clear what regulatory mechanisms other than transcription serve to control late gene expression. T4 Head Morphogenesis When considering mechanisms which may be involved in late gene regulation, it is important to discuss the function of the late gene products in assembly of T4 structural components. It is known for assembly of other supermolecular structures that regulatory mechanisms are in place which coordinate the rate of assembly with the rates of synthesis of the protein components. One example is ribosome assembly, in which groups of ribosomal proteins are encoded by different operons, and some ribosomal proteins repress the translation of the mRNAs that encode them (Nomura et al., 1984). This "feedback" regulation maintains the proper molar ratios of the proteins relative to each other. Another example is the assembly of flagella. The expression of certain flagellar operons depends upon prior assembly of 26 specific flagellar precursor structures. This has led to the hypothesis that transcriptional control of flagellar genes is coupled to the assembly of flagellar structures (Komeda, 1986). T4 head assembly requires the coordinated function of at least twenty structural proteins, assembly factors, and host components. The capsid is assembled in association with the bacterial inner membrane (Laemmli et al., 1970; Simon, 1972) and is initiated by the interaction of T4 gp20 with the membrane (Black and Show, 1983). This is followed by construction of a prohead core or scaffold structure at the membrane attachment site. The core is composed mainly of the products of T4 genes 22, 21, 67, and three small internal proteins (IPs), and it is an obligatory precursor structure around which the major capsid shell protein (gp23) is polymerized. The prohead then undergoes proteolytic processing which results in cleavage of nearly all of its proteins, and the head is filled with DNA to form a mature capsid (Black and Showe, 1983). The major capsid protein, gp23, interacts with other phage and host gene products at an early step of capsid morphogenesis before addition to the core initiation complex. The T4 gene 31 and Ei coli groEL proteins apparently function to solubilize the newly synthesized gp23 and prevent random aggregation of the protein before prohead assembly. Mutations in gene 31 (Laemmli et al., 1970) or in groEL 27 (Georgopoulos et al., 1972) cause gp23 to accumulate in unorganized lumps on the E. ggii inner membrane. Mutations in another host gene, igig, which alter the lipid composition of the inner membrane, result in the same phenotype of lumping of gp23 on the membrane (Simon et al., 1975). These observations suggest that gp23 may bind to the E. ggii inner membrane before the assembly of T4 proheads begins (Binkowski and Simon, 1983). Proper interaction of gp23 with the prohead core is known to be important for determining the length of the T4 capsid. T4 mutants which are deficient in the major core protein, gp22, form aberrant capsid structures called polyheads, which are open ended tubes of polymerized gp23 (Black and Show, 1983). Furthermore, certain mutations in gene 23 lead to production of phage with shorter heads, called petites (Eiserling et al., 1970), or with extra-long heads, called giants (Doermann et al., 1973). These mutations can be suppressed by second-site mutations in genes for core proteins (Doherty, 1982). In order to examine the protein-protein interactions that control head formation, Eiserling et al. (1984) performed experiments to vary the intracellular amounts of T4 head proteins using mixed infections of wild-type and amber mutant phage. They found that the extent of head elongation was profoundly affected by changes in the intracellular ratios of core and capsid proteins. In particular, 28 maintaining the balance between gp22 and gp23 was determined to be essential for proper head formation. It is interesting to note that genes 22 and 23 are adjacent to each other on the T4 chromosome and can be cotranscribed, although gene 23 is more frequently expressed from its own promoter (80% of gene 23 transcripts are monocistronic, as mentioned above). For each new prohead made, approximately 600 copies of gp22 and 1000 copies of gp23 are required (Black and Showe, 1983). Since the ratio of these components is so important in determining the shape of the T4 capsid, it is reasonable to expect that an as yet undiscovered mechanism may exist to coordinate the synthesis of these essential T4 head proteins. MANUSCRIPT BACTERIOPHAGE T4 QQE SITE: A SHORT AMINO ACID SEQUENCE WITHIN THE GENE 23 PROTEIN WHICH CAN AFFECT THE EXPRESSION OF GENES BOTH IN QiE AND IN TRANS Kristin J. Bergsland and Larry Snyder To be submitted for publication to The Journal of Molecular Biology Abstract Gene expression late in infection by bacteriophage T4 can be blocked by the iii protein encoded by thecryptic DNA element, e14. The block is due to the interaction between the iii protein and a short amino acid sequence about one- fourth of the way in from the N-terminus of the major head protein of the virus. Somehow this interaction terminates the transcription (or translation) of the major head protein gene or other genes into which this region has been cloned. The termination can also block the replication of a plasmid containing the region and can block the expression of other genes ig giggg. We propose that the iii protein exposes a normal regulatory mechanism whereby the synthesis of all gene products late in T4 infection is coordinated through the synthesis of the major head protein and the assembly of T4 heads. The regulation is novel in that a nascent polypeptide chain can block the transcription (or translation) of a gene and also trigger the inhibition of other gene expression in trans. 29 30 Introduction Gene expression in bacteriophage T4 is regulated at many levels to ensure the proper temporal appearance of its gene products during development. This complex process involves both phage— and host-encoded gene products and occurs, to a large extent, at the level of transcription although some T4 genes expressed early in infection are also regulated at the level of translation (for reviews see Rabussay, 1983; Brody et al., 1983). Translational regulation can occur through formation of RNA secondary structures which block initiation by ribosomes (McPheeters et a1, 1986) or by binding of a specific translational repressor protein, such as gggg, to the mRNA (Karam et al., 1981). During T4 infection, translation is generally closely coupled to transcription, and this coupling is thought to be important in controlling expression of delayed early genes. It has been proposed that the rate of translation determines whether rho-dependent transcription termination sites between immediate early and delayed early genes are exposed (Daegelen et al., 1982), and this mechanism is thought to be important for modulating phage gene expression according to environmental conditions and the physiological state of the host cell (Brody et al., 1983). With the myriad of diverse controls to which the early genes are subjected, it seems logical that similar complexity might exist late in infection to govern the 31 expression of genes for T4 structural components. However, no mechanisms are known to coordinate the expression of the late genes. As a genetic approach to understanding the regulation of late gene expression, we have analyzed particular mutants of Escherichia coli which block all gene expression at late times after infection by wild—type T4 (Cooley et al., 1979). These bacteria have a mutation in the cryptic DNA element, e14, at 25 min. on the E. ggii K-12 genetic map. These mutations are called iii(Con), and result in overproduction of a 34kD protein, the iii protein (for iate inhibition of 14), which is localized in the bacterial inner membrane (Kao, Gumbs and Snyder, 1987; Kao and Snyder, submitted). The defect in T4 development is specifically on gene expression, since DNA replication is not affected and the T4 genome is not degraded (Cooley et al., 1979). T4 mutations which overcome the defect in gene expression are called ggi, for they grow gn iit(Con). The ggi mutations map within a 40 base pair region about one— quarter of the way into the coding sequence for gene 23, which encodes the major capsid protein of T4 (Champness and Snyder, 1984). However, the complete gene 23 polypeptide is not required for the activity since amber mutations, which truncate the polypeptide, do not cause loss of the activity. Furthermore, ggi mutations are apparently gig-acting; in mixed infections of lit(Con) hosts by wild-type T4 and T4 32 with a ggi mutation, the late genes are only expressed from the DNA with the ggi mutation. There is also a giggg effect on gene expression, since in the mixed infection the rate of protein synthesis is less than 50% of normal even if the phage are added at equal multiplicities (Champness and Snyder, 1982). The ggi site within gene 23 prevents transformation of iig(Con) mutant bacteria when this region of wild-type T4 DNA is cloned into a recombinant plasmid (Champness and Snyder, 1984). Thus, the phenotypes are due to the interaction between the iii protein and this region of the T4 genome. No other T4 genes or activities are required although other host genes could be involved. In this paper, we present evidence that the ii: protein causes a specific termination of translation (or transcription) in gene 23 in the region of ggi mutations. This termination requires a nascent sequence of about 25 amino acids since changing this sequence prevents the termination. Furthermore, termination at this site has other consequences. The replication of a plasmid containing the site abruptly ceases and the expression of other genes is also blocked. We propose that these phenotypes reflect a normal regulation in E. ggii which is somehow exposed by the adventitious interaction of the T4 gol region with lit protein. 33 MATERIALS AND METHODS Bacterial and Phage Strains The bacterial strains used, their relevant characteristics, and the source or reference are listed in Table 1. T4 gene 23 amber mutants were obtained from W.B. Wood and the Boulder T4 stock collection. T4dec8 and the spontaneous gol mutant T4 gol6B are from our laboratory. Mutagenesis Nitrosoguanidine mutagenesis was by the method of Miller (1972). Bacteria containing plasmids to be mutagenized were treated with NTG for two cycles before the plasmid DNA was extracted. Transduction Generalized transduction by P1 was by the method of Miller (1972) with Plvir. The bacteria were washed three times with saline containing 10 mM sodium citrate prior to plating on tryptone plates with tetracycline (10 ug/ml). Determination of G01 and Lit phenotypes The presence or absence of the iig(Con) mutation in E. ggii was determined by streaking a bacterial colony across a streak of wild-type T4 at a titer of about 108/ml. The Gol phenotype of a phage strain was determined by spotting the lysate onto a lawn of BKL155, as in Kao et al., 1987. 34 TABLE 1. Bacterial Strains E. coli Relevant Characteristics Source or References K803 rfm; supE Wood, 1966 B834 rgm; Egp° met- Wood, 1966 BKL155 B834 with lit6 mutation [iii (Con)] this work BK0155 BKL155 cured of e14 [lit°] this work BKL155ggp spontaneous met+ derivative of BKL155 with an amber this work suppressor JM101 A(lac pro) thi, supE, lit+ F' traD36, proAB, lacI Messing, 24 M15 1979 JM101—lit6 JM101 with lit6 mutation [iig(Con)] this work JM101-iii° JM101 lacking iii gene [lit°] this work JMlOlAF JM101 cured of the F-episome pgg' this work JM101-lit6AF JM101-1it6 cured of F-episome pig' this work JM101—lit°AF JM101-1it° cured of F-episome, pro' this work ‘m--_ , , 35 The presence of a ggi mutation in a plasmid clone was detected by marker rescue, as described by Champness and Snyder (1984). Briefly, a culture of 8834 containing the plasmid was spotted onto a lawn of BKL155, and about 107 wild-type T4 phage spotted directly onto the first spot. If the plasmid contains a ggi mutation, T4 recombinants can result which form discrete plaques on BKL155 within the spot. All bacterial growth and incubation was at 30°C. Construction of isogenic lit(Con) and lit0 E. coli strains BKL155 was made by crossing the iigg mutation into 8834 by Hfr conjugation. BK0155 is an isogenic iig° strain which was originally isolated as a survivor of transformation of BKL155 by pLAl (gol’). The strain was cured of pLAl by selecting for tetracycline sensitivity by the method of Maloy and Nunn (1981). The absence of the iii gene in this strain was verified by Southern hybridization (Kao et al., 1987). BKL1555up was isolated by selecting spontaneous met‘ revertants of BKL155 and screening for ability to propagate T4 amber mutant phage. JM101—iigg was obtained through P1 transduction of the iiiE mutation into JM101 from a donor carrying TnlO in fadR, adjacent to the iigg allele. To prepare JM101—iiif, first the TnlO in iggE was transduced from the same donor into BK0155 to get the TetR marker next to the iii° allele. This strain was then used as a donor for P1 transduction of a 36 tetracycline sensitive derivative (isolated as described above), of JM101-iigg to create JM101-iiif. JM101 and its derivatives were cured of the F—episome by selecting survivors of infection with M13Eii, and screening for proline auxotrophy. The resulting strains are designated JMlOlAF, JM101-lit6AF and JM101-lit°AF. Analysis of T4 protein synthesis BKL155 and BK0155 were grown at 30° to 4x103/ml in M9 (5.5g of NazHPO4, 3g of KHZPO4, 0.59 of NaCl, 1.09 of NH4Cl, 0.5% glucose, lmM MgSO4 per liter) supplemented with methionine at 50 ug/ml and tryptophan at 10 ug/ml. Bacteria were infected at a multiplicity of infection of 10 with T4 which had been purified on a CsCl step gradient. Samples of 1 ml of infected cells were labeled with 1 uCi of 14C— labeled amino acid mixture per ml from 30-34 min. after infection. Ice was added at the end of the pulse, the bacteria were spun and resuspended in 0.05 ml of water, and 0.15 ml of sample buffer (2% SDS, 5% -mercaptoethanol, 10% glycerol, 0.1M Tris, pH 6.8 and 0.03% bromOphenol blue) was added. The samples were boiled for 2 min. before loading 25 ul on gels. Gel Electrophoresis The procedures and buffers of Laemmli (1970) were used for SDS-polyacrylamide slab gel electrophoresis. The gels 37 were 10% acrylamide with a 4% stacking gel. After electrophoresis, gels were stained with Coomassie blue to check that total protein per lane was uniform, then destained, dried, and used to expose Kodak X-OMAT AR film for varying lengths of time. The gels containing immunoprecipitated proteins were also treated with Enlightening autoradiography enhancing solution (DuPont) for 20 min. prior to drying. Plasmid Transformation Transformations were performed essentially by the method of Selzer et al. (1978). Cells were grown in tryptone broth (1% tryptone, 0.5% NaCl, 1% casamino acids, thiamine 20 ug/ml) and spread on tryptone plates (tryptone broth plus 2% agar) with 50 ug/ml ampicillin, 50 ug/ml kanamycin, 25 ug/ml chloramphenicol or 10 ug/ml tetracycline when appropriate. When B-galactosidase activity was to be detected, x-Gal (50 ug/ml) and IPTG (0.5 mM) were included. The ability of a plasmid clone of the wild-type ggi sequence to prevent transformation of a iig(Con) strain can be used as an assay for ggi activity. For the standard transformation assay, competent cells were prepared from E. ggii BKL155 [iii(Con)] and BK0155 [iii°] which are isogenic except for the iii gene. Bacteria were grown at 30°C and, following transformation, spread on antibiotic plates and incubated at 30°C overnight. Normal transformation of these 38 strains generally resulted in >1000 transformants per ug of plasmid DNA. A plasmid clone was considered to contain an active ggi region, i.e. able to block transformation of iii(Con), if it yielded at least 1000-fold fewer transformants of BKL155 than of BK0155. A plasmid with a ggi mutation resulted in approximately equal numbers of transformants. Some plasmid mutants had a "leaky" phenotype by this assay, and were considered to have partial ggi activity. These plasmids yield about the same number of transformants of both strains, but the BKL155 colonies were slow-growing and took several days to become visible on a plate. Isolation of DNA fragments from agarose gels DNA digested with restriction enzymes was electrophoresed in 1% agarose gels. The gel was stained with ethidium bromide and cut ahead of the fragment of interest. Nitrocellulose paper NA45 (Schleicher and Schuell) was inserted in the slit and the fragment was electrophoresed onto the paper. The fragment was eluted by heating for one hour at 70°C in 1M NaCl + 50 mM arginine (free base). The fragment was precipitated with 95% ethanol, washed with 70% ethanol, dried and used in ligation reactions. Construction of pA83B17 A recombinant plasmid containing part of gene 23 with 39 the amBl7 mutation was constructed by the following procedure. A T4 multiple mutant which replicates its DNA with cytosine (Snyder et al., 1976) was crossed with T4 with the amBl7 mutation, and the progeny plated on E. ggii B834ga1U56 containing the plasmid pLAl. In this genetic background, only T4 with cytosine DNA will multiply (Runnels and Snyder, 1978) and the plasmid, pLAl, will complement the gene 23 mutation (Jacob et al., 1981). The plaques were spotted on 8834ga1U56 without the plasmid, and lack of growth indicated the presence of the amBl7 mutation. The phage were purified and the HindIII fragment from the N-terminus of gene 23 cloned in pACYC184 as previously described (Champness and Snyder, 1984). The presence of the amBl7 mutation in the recombinant plasmid was confirmed by marker rescue. The bacteria containing the plasmid were grown at 30°C in M98 to 4x108/ml and infected with T4tsH86, which has a temperature sensitive mutation in gene 23 mapping in the amino-terminal end, close to amBl7. After 2 hours, chloroform was added and the lysate diluted and plated on Ei ggii K803 at a restrictive temperature of 42° to select for temperature resistant recombinant phage. Some of these phage had the amber phenotype, confirming that the recombinant plasmid had the amBl7 mutation. Deletion analysis of the gol region (i) Removing sequences from the 3' end: 40 Starting with pUC8H3, DNA sequences 3' to the ggi region were removed by Bal3l digestion. The plasmid was opened at the single EcoRI site in the polylinker, which was adjacent to the HindIII site in gene 23, and treated with Ba13l nuclease using the buffers and conditions specified by the manufacturer. The ends of the deleted plasmids were made blunt by filling in with Klenow polymerase, and phosphorylated EcoRI linkers were added as described in Maniatis et al (1982). The plasmids were prepared from the transformants and checked for deletions by restriction analysis and size determination on agarose gels. Those plasmids which had deletions in the T4 DNA insert were assayed for ability to transform BKL155. The deleted T4 DNA inserts were recloned as HindIII-EcoRI fragments into M13mp19, and the DNA sequence determined in order to locate the endpoint of the deletion. M13 cloning and DNA sequencing were done according to the procedures described by Champness and Snyder (1984) and the Bethesda Research Laboratories M13 cloning/DNA sequencing manual. The HindIII-EcoRI fragments were also recloned into new pUC8 vectors and used for transformation, to make certain that the transformation phenotypes of the plasmids were due to changes in the T4 DNA and not deletions in the vector. DNA fragments for cloning were isolated from agarose gels as described above. 41 (ii) Construction of in-frame subclones of the ggi region in pUC vectors: DNA sequences were removed upstream of the ggi region by subcloning smaller pieces from the 1.1 kb wild—type HindIII fragment using available restriction sites 5' of the ggi region. pUC8PH, which contains a 340bp PstI—HindIII fragment from gene 23 was constructed by isolating the 1.1 kb HindIII fragment from the ggi region and cutting with Pstl, then ligating into pUC8 digested with PstI and HindIII. A modified pUC8 vector, pUC84, was constructed to adjust the translational reading frame for expression of the cloned gene 23 sequences. The BamHI site of pUC8 was cut, filled in by Klenow polymerase and religated, adding 4 bp to the polylinker region 5' to the Pstl and HindIII cloning sites. The same 340 by PstI—HindIII fragment was cloned in this vector to make the plasmid pUC84PH. pUC84PZ1 is a similar construction which contains a 160 bp fragment that extends from the PstI site 5' of the ggi region to the HindIII linker at the site of the Bal3l deletion Al on the 3' side. The T4 DNA insert in this plasmid is in frame with both the iggE ribosome binding site and downstream iggE sequences, which was verified by DNA sequencing. The plasmid pUC84PG4 was isolated after mutagenizing pUC84PZ1 with nitrosoguanidine and recloning the PstI—HindIII fragment into PpUC84. This plasmid is identical 42 to pUC84PZl except that it contains the point mutation go_1NTG4 . Further subcloning was done by isolating smaller fragments from pUC84PZl after electrophoresis on 8% acrylamide gels. These fragments were isolated by cutting slices containing the bands out of the stained gel, placing them in a well of a preparative agarose gel and electrophoresing the DNA onto NA45 paper as above. By this method, a 117 bp PvuII-HindIII fragment was cloned into the HincII and HindIII sites of pUC84 to yield pUC84VZl. A 101 bp RsaI-HindIII piece ligated into pUC12 cut with HincII and HindIII produced pUClZRZl. The plasmid pUC8TZl contains a 72 bp Tan—HindIII fragment cloned in the Ach and HindIII sites of pUC8. The gene 23 coding sequences in these three plasmids were in frame with both the upstream and downstream iggE sequences. To construct the plasmid pUC12PZlR, which translates the ggi region in the "-1" reading frame with respect to gene 23, we first cloned the PstI-HindIII fragment into pUC12. The 5' end of iggE was already in register with the"-l" frame of gene 23 and we put the 3' end in register by adding an 8 bp xhoI linker to the Ball site located to the ggi region. This construct now made colorless transformants due to the nonsense (ochre) mutation in the gene 23 "—1" frame. Bacteria containing the plasmid were then treated with nitrosoguanidine to induce point mutations, in an effort to 43 revert the ochre codon and eliminate the translational block. Plasmids were screened by transforming the mutagenized DNA into the indicator strain JM101-iig°. Several blue colonies were identified by this procedure, and the DNA was sequenced after cloning the PstI-HindIII fragment into M13mp18 to see whether the ochre codon had been changed. One of the plasmids was indeed found to have a single base pair change, a T-A to C-G transition, which changed the ochre codon to a glutamine codon in the "—1" reading frame. The PstI—HindIII fragment was recloned into pUC12, and this plasmid is pUC12PZ1R. Construction ofypSKS-PZl and pSKS-PG4 Translational fusions of the ggi region with the entire iggE gene in the vector pSKSlOS were made using the 160 bp PstI-HindIII fragments from pUC84PZ1 and pUC84PG4. Since pSKSlOS only has unique cloning sites for BamHI and HindIII, the 160 bp fragments were first cloned in pUC8, and then BamHI-HindIII fragments were prepared from these plasmids and cloned into pSKSlOS, to make pSKS—PZl and pSKS-PG4. Although these fragments were not expected to be in frame on the 5' end with iggE sequences, there was clearly a significant amount of iggE expression, since transformants of E. ggii JMlOlAF were dark blue on X—Gal media. Attempts to put the 5' end in frame by filling in the BamHI site were unsuccessful, probably because high level expression of - 44 galactosidase is toxic to the bacteria. pSKSlOS itself was very unstable in this regard, and was prone to accumulating insertions and deletions in the igg region. Nevertheless, DNA sequencing of the EcoRI-HindIII fragments cloned in M13mp19 confirmed that the constructions were identical to each other except for the ggi point mutation, and that the gene 23 sequences were in frame with lacZ on the 3' end. Analysis of B-galactosidase synthesis B-galactosidase activity was assayed by hydrolysis of ONPG as described by Miller (1972). Cells were washed once in 0.1M sodium phosphate buffer, pH7.0 and lysed by sonication. For labelling with 35S-methionine, cells were grown in M9 medium with 0.4% glycerol, thiamine (20 ug/ml) and all amino acids except methionine at 50 ug/ml, plus 25 ug/ml ampicillin and 0.5 mM IPTG. The bacteria were grown to 5x108/m1 under the conditions indicated for each experiment. Then 3 ml of culture was labelled with 30 uCi 35S-methionine per ml for 15 minutes. Cells were washed once with cold 50 mM Tris, pH 8.0 and resuspended in 50 ul of a solution containing 1% SDS, 50 mM Tris, pH 8.0 and 1 mM EDTA, then boiled for 3 min. Immunoprecipitations were done essentially according to the method of Hall et al. (1983). Briefly, 30 ul of the SDS— solubilized protein was added to 960 ul of cold Triton buffer 45 (2% Triton X-100, 50 mM Tris, pH8.0, 150 mM NaCl and 0.1 mM EDTA) and microfuged for 10 minutes. One ug of mouse monoclonal anti-B-galactosidase (Promega Biotec) was added to the supernatant and left overnight at 4°C. To this mixture was added 50 ul of protein A-Sepharose (50% solution) and this was incubated on ice with occasional mixing for one hour. The immune complexes were then pelleted in the microfuge for 5 min., washed twice with cold Triton buffer and once with 10 mM Tris, pH8.0. The final precipitates were extracted with 50 ul sample buffer by boiling for 5 minutes. The samples were microfuged for 3 min. before loading 40 ul on an SDS—polyacrylamide gel. Materials [d-32PJdATP was from ICN Radiochemicals (>600 Ci/mmol). 35S—methionine was from Amersham (>800Ci/mmol) and the 14C- labelled amino acid mixture was from New England Nuclear. The M13 cloning/ dideoxy sequencing kit, X-Gal, IPTG, pUC8 and pUC9 were products of Bethesda Research Laboratories. The antibiotics and ONPG were from Sigma Chemical Co. Restriction enzymes and Bal31 nuclease were from International Biotechnologies, Inc. and were used with the supplied buffers. T4 DNA ligase and pUC12 were purchased from Boehringer Mannheim Biochemicals. DNA polymerase I large fragment (Klenow enzyme) and synthetic DNA linkers were from New England Biolabs, Inc. 46 Results Translation over the region of 901 mutations is required for activity of the site Genetic mapping and DNA sequencing have shown that ggi mutations lie within a 40 base pair region about one-fourth of the way in from the amino—terminal end of gene 23. Preliminary experiments had suggested that the phenotype was not due to the polypeptide translated in the gene 23 frame because amber mutations in gene 23 did not alter the phenotype. However, since it was possible that only part of the gene 23 polypeptide was required, we decided to reinvestigate the effect on T4 gene expression of gene 23 amber mutations, both upstream and downstream of the region of ggi mutations. Late protein synthesis was analyzed in non—amber—suppressing bacteria infected with four gene 23 amber mutants. Two of the amber mutations lie upstream of the ggi region, amHll and amBl7, and two are downstream, amB272 and amE389 (Fig. 1). If translation of the ggi region is required for its activity, we would expect the two upstream mutations to allow the synthesis of all the T4 proteins except the product of gene 23 in the iii(Con) host. The behavior of the downstream mutations would depend upon how much of gene 23 must be translated for the activity. We Show the results for two of the mutations, amBl7 and amB272, in Fig. 2. Infection of lit(Con) bacteria with T4 having the Fig. 1. 47 Genetic and restriction map of the gene 23 region of bacteriophage T4. The locations of gene 23 amber mutations and relevant restriction sites are shown. The position of the 1.1 kb HindIII fragment which was previously determined to block transformation of iig(Con) bacteria is indicated (Champness and Snyder, 1984). Transcription is from left to right. Abbreviations: H, HindIII; D, Dral; P, Pstl; V, PvuII; R, Rsal; T, Tan; B, BalI. 48 I mh¢>m numb- nv. P... D P 0.5mm. Q? 0N ocom L. _ NN «com 6883um "' ZZZBWD '4‘ ZLQWD "" uno —[ 49 gene 23 amber mutation located upstream of the gol region resulted in normal patterns of late gene expression when compared with infections of the iii? strain. The same result was obtained with the amber H11 mutation (data not shown). Therefore, these N—terminal amber mutations have the same phenotype as T4 ggi mutations except that they do not form plaques because of the amber mutation in gene 23. In contrast, infections of iiE(Con) E. ggii with phage with the amber mutation located just downstream of the ggi region showed striking inhibition of protein synthesis, behaving like wild—type T4 in blocking late gene expression. The same result was obtained with amE389 (data not shown). We conclude from this experiment that translation as far as the ggi region is required for its activity in preventing gene expression in a ii:(Con) host, but that translation need not extend much past this site. This conclusion is also supported by the results of a transformation experiment using a recombinant plasmid, pA83Bl7, containing the ggi region with the amber mutation amBl7. As mentioned above, the presence of a wild-type gene 23 sequence in a plasmid prevents transformation of the plasmid into a ii§(Con) recipient, and this can be used as a functional assay for the activity of the site. The pA83B17 plasmid yielded about a thousand times fewer transformants of an amber—suppressing iii(Con) strain, BKL155ggp than of the isogenic non-suppressing strain BKL155. In other words, Fig. 2. 50 T4 late protein synthesis after infection of lit(Con) E. coli by gene 23 amber mutants. An autoradiogram of T4 proteins labeled 30-34 min. after infection is shown. The proteins in lanes 1-2 and 3-8 were from separate experiments, both performed in the same way as described in Materials and Methods. The bacteria did not contain an amber suppressor and were either iii(Con) (lanes 1, 3, 5, 7) or lit° (lanes 2, 4, 6, 8). Lanes 1-2, T4amBl7; 3-4, T4amB272; 5-6, T4goll6B; 7-8, wild-type T4. 51 Figure 2 51a Table 2 Plasmids Relevant Characteristics Transformation Source or of BKL155 Reference pLAl pBR322 with a 3.5 kb EcoRI - Jacobs et clone of gene 23 fmn wild al., 1981 type T4; Tetr pACYC184 Camr Tetr ; HindIII site in + Chang & Fit gene Cohen, 1978 pA83 pACYC184 with the 1.1kb HindIII - (harmless & fragment from wild—type T4; Can? Snyder, 1984 pA83Bl7 pA83 with amber B17 mutation + this work pUC8H3 pUC8 with wild-type T4 1.1kb Hind III - this work pUCBI-I3A1 Bal31 deletion; has ntl-397 - this work of gene 23 pUC8H3A60 Bal3l deletion; has nt 1-356 i this mrk of gene 23 pUC8H3A20 Bal3l deletion; has nt 1-313 + this work of gene 23 pUC8PH 340 bp wild-type PstI—HindIII of — this work gene 23 pUC84 pUC8 with BamHI site filled in + this work (4bp added) pUC84PH 340 bp wild-type PstI-HindIII; — this work in frame with lacz r’os pUC84PZ1 160 bp wild-type PstI-deletionAl; - this work lacZ fusion pUC84PG4 pUC84PZ1 with go_lN'IG4 mutation; + this work lacZ fusion pUC84VZl 117 bp wild—type PvuII-Al of gene - this work 23 pUC12RZl 101 bp wild—type RsaI-Al of gene + this work 23 pUC8TZ1 72 bp wild-type Tan-Al of gene 23 + this work pUC12PZ1 160 bp Pst- Alzgzs "—1" reading frame pUC12PZIR TAA codon in pUC12PZl reverted by + this work point mutation; lacZ fusion pSKSlOS polylinker from pUC8; lacIPOZYA + Shapira et a1. , 1983 pSKS-PZl pSKSlOS with PstI-HindIII frtm - this work pUC84PZl; lacZ fusion pSKS—PG4 pSKSlOS with PstI—HindIII fran + this work pUC84PG4; ggi mutant lacZ fusion nt = nucleotide rbs = ribosome binding site 52 stopping translation before the ggi region at the amBl7 mutation in the non-suppressing host permits transformation of the plasmid into a iig(Con) strain. Thus translation of the g9; region is required both to block T4 gene expression and to prevent transformation of iii(Con) bacteria by plasmids containing the region. Defining the limits of the 901 region The plasmid transformation assay can be used to determined the minimum DNA sequence required for activity of the ggi region. It was previously reported that the 1.1 kb HindIII fragment from wild—type T4 which contains part of gene 22 and the amino-terminal one—third of gene 23 including the ggi region (Fig. 1), had all the sequences necessary to prevent transformation of a iii(Con) recipient (Champness and Snyder, 1984). Starting with a plasmid clone of this fragment, pUC8H3, DNA sequences 3' to the ggi region were removed by Bal3l deletion and 5' sequences were removed by subcloning as described in Materials and Methods. Deleted plasmids which were able to transform a iii(Con) host had lost sequences essential for activity of the region. Plasmid clones of the ggi region and their transformation phenotypes are listed in Table 2. A deleted plasmid with the deletion Al which ends 47 bp downstream from the site where the 3'-most ggi mutation has been identified (Fig. 3), still blocked transformation. A deletion (A20) 53 which removes all but 3 bp of the 40 bp ggi region inactivated the site, and the plasmid transformed iii(Con) bacteria efficiently. Two other deletions which fall 6 bp and 9 bp away (A60 and A56) from the site of the 3'-most ggi mutation are particularly interesting. These plasmids seem to retain only partial gol activity, since they were "leaky" in the transformation assay. Rather than preventing transformation entirely, these plasmids yielded about as many transformants of a iig(Con) strain as of a iii° strain, but the iig(Con) colonies were small and slow—growing and took several days to become visible on a plate. These mutations of intermediate phenotype cause a reduction of function of the region and probably remove sequences which are important, but not crucial, for activity. The deletion, A60, probably comes close to defining a 3' limit of the sequences required since it removes sequences to only 6 base pairs from the 3'— most ggi mutation. We anticipated that it would prove more difficult to delete sequences 5' to the ggi region, since these must leave a functional ribosome start site in frame with the gene 23 sequence. Rather than construct random deletions, many of which would be out of frame, we decided to use restriction sites upstream of the ggi region and subclone fragments such that they would be in frame with the iggE ribosome initiation site of the pUC vectors. This sometimes entailed filling in a restriction site in the vector to change the reading frame. Fig. 3. 55 DNA and amino acid sequences of gene 23 in the gol region showing mutational changes and deletion endpoints. The DNA sequence is written as that of the gene 23 message. Nucleotide numbers above the ends of line and amino acid numbers below are from Parker et al., 1984. Restriction sites used for subcloning are indicated. The ochre codon TAA in the "—1" reading frame is boxed, and the T to C transition in bold-face indicates the mutation which changes the ochre but not the proline in the gene 23 frame. Mutations which are not designated ggi do not confer on T4 phage the ability to grow on iii(Con) bacteria, but share other phenotypes with ggi mutations. 56 Pqu Rsal l"__—'fi l'_'_fi 306 CCAGCTGTTATGGGTATGGTACGTCGTGCT 3 Pro Ala Val Met Gly Met Val Arg Arg Ala 102 golNTG4 goIHA1;2 ""310 gol2AP13 'r CA20 A c 307 336 T f Tamil 1 . . r-——I ATTCC CCTGATTGCTTTCGATATTTGT Ile Pro Asn Leu Ile AlaPhe Asp IleCys 103 112 ser pro asn thr amNTG1 golSB NTG2 ATG NTG-6 C A60 A56 \l f 1 BaII 7GGTGTI'CAGCCGATGAACAGCCC‘GAC'TGGC366 aGlyValGlnProMetAsnSerProThrGly 33 11 122 Iys am arg thr A1 367— [397 CAGGTATTCGCACTGCGCGCAGTATATGGTA 3(:‘nanalPheAl aLeuArgAlaValTyrGIV 12 132 Figure 3 57 We found that a subclone of the PstI to HindIII fragment of gene 23 (see Fig. l) retained the activity of the ggi region, since it blocked transformation of the iig(Con) recipient. However, a clone from the Tan site which falls almost in the middle of the ggi region (pUC8TZ1), and eliminated 16 bp of the 40 bp region (Fig. 3) did permit transformation into BLK155 so did not have the activity. A fragment starting from the PvuII site which is 30 bp upstream of the location of the 5'-most ggi mutation (pUC84VZ1) retained the activity of the site. However, a fragment from the RsaI site, between PvuII and Tan and just 16 bp smaller than the PvuII fragment (pUC12RZl), did not have the activity. Thus, the 5' limit of the ggi region appears to lie between the PvuII and RsaI sites, since removing this sequence eliminates the ability of a clone to prevent transformation. In summary, sequences can be removed from the 3' end of gene 23 up to 6 bp from the 40 bp region of ggi mutations, and from the 5' end, at least as far as the PvuII site 30 bp upstream, but not as far as the RsaI site, and a functional ggi sequence is still retained. This demonstrates that, at most, 75 bp of DNA are necessary for the activity of the ggi region. 58 A specific amino acid sequence of gp23 is required for the phenotype Although the experiments above showed that translation over the ggi region is required to prevent transformation and to block T4 gene expression, it is possible that merely the act of translating per se rather than a specific amino acid sequence from the ggi region is required. To determine whether a specific amino acid sequence from the ggi region is required, we compared the transformation phenotypes of two plasmids containing the same DNA sequence but which encoded different polypeptide products from the ggi region. pUC84PZl has a 160 bp PstI-HindIII fragment containing the wild—type ggi region which is translated in the gene 23 reading frame. A second plasmid was constructed by cloning the same fragment into pUC12 (called pUC12PZl), in which translation will be in the "-1" reading frame with respect to gene 23. However, the "—1" reading frame in the cloned gene 23 sequences is interrupted once by an ochre nonsense codon. We eliminated the ochre codon in the pUC12 clone by nitrosoguanidine mutagenesis, to make pUC12PZlR, as described in Materials and Methods. Conveniently, the mutation which changed the ochre codon was silent in the gene 23 reading frame. The two translational fusions, one in the gene 23 reading frame and the other in the "—1" frame, were then tested for ggi activity by the transformation assay. pUC84PZ1 did not transform BLK155, while pUC12PZlR did 59 transform efficiently and behaved as a ggi mutant. Thus, simply translating through the ggi region does not seem to be sufficient for the activity, but the specific amino acid sequence in the gene 23 reading frame is required. However, one could argue that the addition of the XhoI linker and/or the point mutation which changed the nonsense codon could inactivate the region. Consequently, we examined the effect of adding an XhoI linker to the Ball site of pUC84PZl, and found that the phenotype was unchanged. We also recloned the PstI—HindIII fragment with the changed ochre codon from pUC12 into pUC84 to restore the gene 23 reading frame. This fragment, which had lost the activity in pUC12, now regained the activity in pUC84, indicating that this single base pair change was not itself acting as a ggi mutation. Thus, the amino acid sequence in the gene 23 reading frame is required for gol activity. The gol region inhibits expression of downstream sequences Previous experiments demonstrated that ggi mutations are gig-acting for expression of gene 23 in mixed infections of a iig(Con) host by gol+ and ggi mutant T4 (Champness and Snyder, 1982). It is not obvious how the ggi region could affect gene 23 expression in gig, considering that it is located a quarter of the way into gene 23 and encodes a specific amino acid sequence required for its activity. One possible explanation is that in the presence of excess lit 60 protein, the wild-type ggi site interferes with gene expression such that transcription or translation cannot proceed past this point. To test this, we again used translational fusions with iggE. In Fig. 4, we show a schema of the fusion plasmids. Basically, the ggi region of T4 DNA is cloned into the pUC vector such that the gene 23 reading frame is in register with the iggE coding sequences on both ends. The translation will initiate at the iggE ribosome initiation site and translate through the T4 sequences into the downstream iggE sequences. If translation, or transcription, stops in the ggi region, the iggE polypeptide will not be made and the bacteria harboring the plasmid will make a colorless colony on X-gal plates. If translation or transcription proceeds through the site, the colony will be blue. The two plasmids used in this experiment were the wild-type gol-lacZ fusion, pUC84PZ1, and a derivative, pUC84PG4, which was identical except that it contained the point mutation ggiNTG4, which has been described previously (Champness and Snyder, 1984). We examined the effect of the ggi region on expression of downstream iggE sequences in the presence or absence of the iii protein by transforming these plasmids into three E. ggii indicator strains which were isogenic except for the iii gene: JM101 has wild-type e14 so makes normal amounts of lit protein; JM101—lit6 has a lit(Con) mutation in e14 which causes the overproduction of lit protein; and JM101—lit° is 61 Fig. 4. Schema of structure of translational fusions with T4 gol sequences fused to lacZ. T4 DNA containing the ggi region was cloned into the polylinker of the pUC plasmids or pSKSlOS, internal to the lacZ coding sequence. The 5' and 3' ends of the insert are in frame with the lacZ sequences on either end, and a fusion protein is produced under the control of the igg transcriptional and translational regulatory sequences. 62 IacZ Figure 4 63 cured of e14 so has no ii: protein. Transformants of JM101—iii° with either the wild-type or the ggi mutant fusions were dark blue on x-Gal plates. However, there was a striking difference in blue color between the wild-type and ggi mutant transformants of JM101 (iii*). While the ggi mutant fusion gave dark blue transformants, the transformants with the wild—type fusion were only faintly blue. The two types of colonies were approximately the same size and contained approximately the same amount of intact plasmid DNA when plasmid minipreps were examined on agarose gels. The wild-type fusion gave many fewer transformants of JM101-iiig, as expected, and these few were very small and generally colorless. The ggi mutant fusion produced normal-sized dark blue transformants of JM101—iigg, as with the other strains. These results strongly suggest that there is an interaction between the iii gene product and the ggi region which prevents the expression of downstream sequences. The ggi mutations overcome this effect and also somehow alleviate the effect on plasmid transformation. From this experiment, it is also evident that the ggi region exerts some effect even in the presence of wild-type levels of the iii protein, and not just when iii protein is overproduced as in a iii(Con) host. It is possible that the relative inability to express the lacZ sequences in the above experiment is due to a 64 preferential inhibition of iggE expression and not to a gig- effect of the wild-type T4 sequences on the same plasmid. If so, then other iggE expression in the same cell should be likewise inhibited. However, the wild-type ggi sequence does not inhibit iggE expression in the chromosome or from a compatible plasmid in cotransformants (data not shown). It is only when the ggi site is cloned in gig, as in pUC84PZ1, that inhibition of iggE_gene expression occurs in the presence of wild-type levels of lit protein. Inducing mutations in the 901 region of plasmid clones We reasoned that since the golNTG4 mutation allowed expression of the gol-lacZ fusion in the presence of the lit protein and therefore made blue colonies on X-Gal plates, we might be able to use this property to identify new ggi mutations. A similar approach was employed in earlier work from this laboratory, which showed that ggi mutations could be induced in plasmid clones and selected by ability to transform a iii(Con) recipient (Champness and Snyder, 1984). It was predicted that mutations induced in the plasmid might define a broader class of ggi mutations than would be found in the T4 phage, since mutations selected in the phage must leave the function of the gene 23 protein intact, but there is no such requirement for mutations selected in plasmids. Several plasmid mutants were found this way which could transform lit(Con) recipients, but most did not confer the 65 Gol phenotype on T4 in marker rescue tests. However, the exact nature of these mutations was never determined since it was difficult to identify where they lay within the large fragment in which they were detected. Consequently, the advantage of using the ol—iggE fusion in pUC84PZl to identify new mutations is that the T4 DNA insert is small (160 bp) and can easily be cloned into M13 vectors for DNA sequencing to precisely locate mutational changes. This strategy also permits us to ask whether mutations which allow the expression of iggE are also ggi mutations by other criteria, that is by allowing plasmid transformation into iii(Con) bacteria or T4 growth on a iig(Con) host. For this experiment, we mutagenized cells carrying pUC84PZl with nitrosoguanidine, and the plasmids were prepared and used to transform JM101. The dark blue colonies were picked and retested for the presence of a mutation in the plasmid. However, in most of these the chromosome had been cured of e14 by the criterion that the plasmid still did not give transformants of iig(Con) recipients. Furthermore, when e14 was reintroduced by P1 transduction selecting for tetracycline resistance on a closely linked Tn10 transposon, the colonies were again only faintly blue. Since the dark blue transformants were much more likely to have lost the e14 DNA element than to have a mutation in the plasmid, we revised our approach to select for plasmid mutants. The NTG—mutagenized pUC84PZ1 plasmids were first 66 transformed into the iig(Con) strain, BKL155, and these transformants, which could be either iig° or contain a ggi mutant plasmid, were screened for maintenance of the iig(Con) allele by cross-streaking on T4’. Plasmids were prepared from those that were still iig(Con) and screened for blue color in JM101 and for ability to confer the ggi phenotype on T4 by marker rescue. The PstI-HindIII fragments from these plasmids were also cloned into M13mp18 for DNA sequencing. Each of the 13 plasmids chosen for further analysis was found to have a single base pair change within the 40 bp ggi region of the T4 DNA insert. Only five of 13 conferred the G01 phenotype on T4 by marker rescue and these all gave blue transformants. They were all found to have the same base pair change as the previously identified ggi mutants NTG4, HA1 and HA2, a C-G to T—A transition at nucleotide number 310 of gene 23 (Fig. 3). Four others which gave faintly blue transformants had a C-G to T-A transition at position 343, which changed a glutamine codon to an amber nonsense codon. Another mutation, NTGZ, altered the same base pair by a C-G to A-T transversion, changing the glutamine codon to a lysine codon. The mutation NTG6 caused an A—T to G-C transition in the same codon at position 344, changing it to encode arginine. Finally, two other mutations changed the G-C at position 328 to an A-T, causing the aspartate codon to now code for asparagine. All of these mutations except the amber mutations caused 67 transformants of JM101 to form dark blue colonies on X-Gal plates. Thus, there is an absolute correlation between ability of mutations to allow transformation of iig(Con) bacteria and to allow expression of sequences downstream of the ggi region. It is expected that the amber mutations would allow less expression of the downstream iggE sequences than the other mutations. They must be suppressed and the surrounding "context" sequences suggest they will be suppressed poorly by gng (Bossi, 1983). However, these amber mutations do confer a significantly bluer color than the wild-type sequence in spite of their poor suppression. This is particularly interesting considering that gng inserts glutamine, and so restores the wild-type amino acid sequence. All of the mutations isolated in this way are ggi mutations by the criteria that they permit transformation of iii(Con) recipients by the plasmid and they stimulate expression of the downstream iggE sequences in the fusion in the presence of iii protein. However, only those at nucleotide 310 conferred the G01 phenotype on T4 bacteriophage. Because we have defined ggi mutations as those which permit T4 to grow gn iit, we have only referred to the others by their mutation number and have not called them gol mutations (see Fig. 3). 68 The 901 region can affect both gene expression and the replication of a plasmid containing the site In an effort to quantify the effect of the ggi region on iggE expression, we made translational fusions with the entire iggE gene in the vector pSKSlOS (Shapira et a1, 1983). This plasmid contains the polylinker and lacIPO region from pUC8 attached to the whole igg operon rather than just the £29! d-peptide sequences. The advantage of this system is that the entire B-galactosidase is made and the enzyme can be more easily detected. By using antibody specific for the B -galactosidase protein, we can ask whether the B - galactosidase is made, but is somehow not active, or if the synthesis of the protein is blocked. The bacterial hosts used for the pSKSlOS plasmids were all JM101-derived strains (such as JMlOlAF) which had been cured of the F-episome carrying iggE sequences to avoid recombination between iggE sequences in the plasmid and in the episome. The wild-type and ggi mutant fusions, pSKS—PZl and pSKS- PG4, were constructed as described in Materials and Methods. These clones were expressed at a low level (see above) which proved to be a useful feature, since pSKS-PZl could be maintained in a iig(Con) strain without the severe toxicity observed with the pUC clones. But pSKS-PZl clearly responded to excess amounts of the ii: protein when transformed into JM101-iiggAF carrying a iig(Con) mutation. pSKS-PZl transformants were small with faintly blue centers, while the 69 pSKS-PG4 colonies were larger and dark blue. In the experiment in Figure 5, the two strains JM101-iii°AF and JM101—iigEAF containing either pSKS-PZl or pSKS-PG4, were grown to exponential phase at 30°, then B-galactosidase activity was assayed and the proteins labeled with 358— methionine were precipitated with anti-B-galactosidase. If the ggi amino acid sequence inhibited the enzymatic activity and not the expression of the fusion protein, than the B- galactosidase protein could be visualized on a gel after immunoprecipitation. In this experiment, there was a hundred-fold less B—galactosidase activity in the JM101- iiEQAF/pSKS-PZl culture compared to JM101-iigg F/pSKS-PG4, and the clone with the wild-type sequence made undetectable amounts of protein which reacted with antibody to B- galactosidase. When we examined the plasmids in each of the transformed strains, we discovered that the copy number of the pSKS-PZl plasmid with the wild-type ggi sequence was much lower in the iii(Con) strain. The plasmid with the ggi mutant sequence showed no such drop in copy number nor did the pSKS-PZl plasmid in the iii° strain (see Fig. 5). This drop in plasmid copy number is only apparent in the presence of higher amounts of ii: protein, since we did not observe it with the pUC plasmid clones in JM101 with wild-type amounts of ii: protein. The effect on the plasmid acts in gig because the copy number of a cohabitating, compatible plasmid Fig. 5. 7O ExpreSsion of gol—lacz fusions and plasmid replication in lit(Con) bacteria at 30°C. A. An autoradiogram of immunoprecipitated B— galactosidase labeled for 15 min. with 358- methionine. ' B. An ethidium bromide stained 0.8% agarose gel after electrophoresis of HindIII-digested plasmid DNAs. Plasmids were prepared from each culture by an alkaline lysis miniprep procedure at the time the proteins were labeled. Lanes 1, JM101-iii°£F/pSKS-P21; 2, JM101-iiifA F/pSKS-PG4; Lanes 3, JM101-lit6AF/pSKS-PZl; 4, JM101-lit6AF/ pSKS-PG4. 71 Figure 5 72 which does not contain the T4 ggi sequence was not similarly affected (data not shown). In this particular experiment, the drop in plasmid copy number seems sufficient to explain the defect in B— galactosidase synthesis. However, from our results with the pUC plasmid clones, in which there was no obvious effect on plasmid copy number, we thought there was also a direct inhibition of the expression of downstream iggE sequences. However, to prove this it was necessary to separate the effect on plasmid copy number from the effect on B- galactosidase synthesis. To separate these two effects we took advantage of the temperature dependence of the phenotypes. At 42°C, much more lit protein is required to block T4 development than at 30°C (Kao and Snyder, submitted). We reasoned that if we grew the cells at 42°C, both plasmid copy number and B—galactosidase synthesis would be approximately normal. If we then shifted the temperature down to 30°C, both plasmid replication and B— galactosidase synthesis might be blocked. However, assuming that the plasmid is not degraded, its copy number should drOp slowly as it is diluted out by subsequent cell growth, while the block in B—galactosidase synthesis should be immediately apparent. In the experiment shown in Fig. 6, pSKS-PZl and pSKS— PG4 were transformed into JM101—iii°AF and JM101—iiiEAF at 42°. As expected, on X—Gal plates the pSKS-PZl transformants 73 of JM101-iigEAF at 42° were large and dark blue, just as the pSKS-PG4 colonies. We maintained the transformants at 42°C and grew them to log phase at 42°C before shifting the temperature to 30°C. We then labeled the proteins with 358- methionine, as before, and immunoprecipitated the proteins and antibody to EPgalactosidase. We also prepared the plasmids to determine their copy number. Now, the copy number of pSKS—PZl in JM101-iigEAF was only about a factor of two lower than pSKS-PG4, but pSKS-PZl made at most one-tenth as much Efgalactosidase in the lit° strain. We conclude that the interaction of the lit protein with the wild—type gol sequence results in a cis inhibition of both gene expression and, under certain conditions, plasmid replication. Discussion Previous experiments had suggested that ggi mutations define a gig—acting site which can prevent all T4 gene expression late in infection of a ii§(Con) host. The evidence for this was that ggi mutations, which overcome the effects of the site, are gigeacting for the expression of gene 23 and possibly other closely linked genes. There also seemed to be a igggg-effect due to the site since in mixed infections with equal multiplicites of wild-type and ggi mutant phage, the rate of late protein synthesis was less than one-half of normal. Even though gol mutations lie ext Fig. 6. 74 Plasmid replication and expression of gol-lacz fusions in lit(Con) bacteria first grown at 42°C, then shifted to 30°C. Plasmid transformants were maintained at 42°C prior to use. For the experiment, the cultures were grown at 42°, then shifted to 30° for one hour before labeling protein and preparing plasmid DNA. A. An autoradiogram of immunoprecipitated B' galactosidase labeled for 15 min. with 3SS—' methionine. B. An ethidium bromide-stained 0.8% agarose gel containing HindIII-digested plasmid DNAs. Lanes 1, JM101-lit°AF/pSKS-PZ1; 2, JM101- lit°AF/pSKS-PG4; 3, JM101-lit6AF/pSKS-PZl; 4, JM101-1it6AF/pSKS-PG4. 75 Figure 6 76 within gene 23 it was concluded that the phenotypes were not caused by an altered gene 23 protein since amber mutations in gene 23 did not alter the phenotypes. The site was known to be active in the absence of the rest of the T4 genome since plasmid clones of this region of the T4 genome prevented transformation of iig(Con) recipients (Cooley et al., 1979; Champness and Snyder, 1982; Champness and Snyder, 1984). In this paper, we show that a short amino acid sequence of about 25 amino acids of the gene 23 protein is required for the phenotypes. Presumably, the interaction between the iii protein and this short amino acid sequence causes the gig-inhibition of the expression of gene 23 and the giggg inhibition of the expression of other genes in E. ggii. The interaction can also cause a gig inhibition of the replication of a plasmid in which the site resides. We have used iggE fusions in which the T4 ggi region is cloned internal to, and in register with, iggE coding sequences to show that the ggi region can cause a gig~ inhibition of the expression of downstream sequences. We think that the interaction between the gol region and the lit protein causes a specific termination. At present, we do not know if the primary effect is a transcriptional or translational termination, although the effect on plasmid replication is easier to explain if the termination is transcriptional. The RNA polymerase may stall at the site and impede the passage of replication forks or the nascent 77 ggi polypeptide may drag the plasmid to a region of the cell not suitable for replication. Somehow, the termination also causes a giggg-acting inhibition of other gene expression. We think this inhibition is specific for gene expression because other cellular processes are normal. Perhaps, the stalled RNA polymerases or ribosomes cause a giggg-acting inhibitor to be generated in analogy to the generation of guanosine tetraphosphate by stalled ribosomes. The T4 ggi mutations change the peptide sequence so that it no longer interacts with iii protein. The termination no longer occurs and neither do the other phenotypes which are a direct consequence of the termination. At present, we do not know if these phenotypes reflect a normal regulation of T4 gene 23 synthesis, somehow exposed, either deliberately or adventitiously, by the presence of the e14 iii protein, or whether they only occur in the presence of ii: protein. It is conceivable that the iii protein of e14 is part of a defense mechanism directed against T-even coliphages. The iii protein may recognize a short amino acid sequence common to the major head proteins of these phages and trigger the phenotypes, thereby preventing phage production and protecting neighboring cells which also harbor e14. It is intriguing that e14 also contains a restriction endonuclease gene directed against unglucosylated hydroxymethylcytosine (E. Raleigh, pers. commun.) which may also be part of such a defense mechanism. 78 Even if the interaction with ii: protein is deliberate, the phenotypes which are exposed may be part of a normal regulatory mechanism. Such a regulatory process might allow the coordination of all gene expression late in infection with the synthesis of gene 23 protein and its assembly into T4 heads. The synthesis of gene 23 protein may normally pause in the ggi region and only continue if the protein is incorporated properly into a T4 head. In this way, the gene 23 product may be fed directly into the growing head as it is synthesized. If the assembly is delayed, for some reason, the pause may be extended, and a giggg-acting inhibition of other gene expression may delay cell lysis and the synthesis of other phage components. In this respect, it is intriguing that T4 head assembly is thought to occur on the inner membrane (Laemmli et al., 1970; Simon, 1972), and ii: is an inner membrane protein. Perhaps the iii protein mimics or somehow otherwise interferes with the normal membrane receptors for T4 head assembly. If the ggi region plays a role in normal phage development, we might expect mutations which inactivate the site to be lethal to the phage. These would be difficult to distinguish from mutations which alter the conformation of the gene 23 protein and thereby preclude the formation of viable heads. In fact, this adequately explains why some mutations isolated in the ggi region in a plasmid do not confer the Col phenotype on T4 bacteriophage. However, it is 79 possible that some of these mutations may be inviable in the phage because they inactivate the function of the ggi region. The amber mutations in the ggi region, represented by amNTGl, are particularly intriguing in this respect. In the iggE fusion clones, the amNTGl mutation allows transformation of iig(Con) recipients even with amber suppressors gng or gng. It also allows more expression of downstream iggE sequences than the wild—type ggi region in JM101 with wild-type e14 and the supE44 mutation. Since supE inserts glutamine for the amber codon and this is the amino acid in the wild-type protein, the amNTGl mutation seems to prevent the termination without changing the amino acid sequence of the protein and must act by changing the RNA or the DNA. Termination sites are often associated with hairpin loops in the RNA, and there is a short dyad symmetry in the 3' part of the ggi region (see Fig. 7). Some of the ggi mutations, including amNTGl, shorten the dyad symmetry. Only one of these, ggi6B, is viable in the phage and this destroys an external base pair so may not totally prevent formation of the hairpin. It is possible that these mutations act to prevent termination by destroying the hairpin, rather than by changing the amino acid sequence, and that the termination is required for normal T4 head assembly. Whether or not these phenotypes reflect a normal regulation, the fact that they can occur suggests that mechanisms are in place to regulate the translation or 80 Fig. 7. A possible secondary structure in the gol region of gene 23. A region of dyad symmetry from nucleotide 339- 355 could form a 6 bp hairpin structure. Some mutations in the ggi region would be predicted to disrupt this structure, and most of these are inviable in the phage. '81 000 D > 240m QAIII>._..||YO 02mm 2: 2.8... ._.ll.loo 23» >\\a.> 00323 a 0 ._.> e no >4104004>00000>04 was new 20:; V 82 transcription of a gene retroactively through interaction with the nascent peptide chain. Furthermore, the translation of one gene can influence the expression of other genes and even the replication of the DNA in which the gene is located. We think it likely that similar regulatory mechanisms are used in other situations. REFERENCES Bossi, L. 1983. Context effects: Translation of UAG codon by suppressor tRNA is affected by the sequence following UAG in the message. J. Mol. Biol. 164: 73-87. Brody, E., D. Rabussay, and D.H. Hall. 1983. Regulation of transcription of prereplicative genes, p. 174-183. In C.K. Mathews, E.M Kutter, G. Mosig and P.B. Berget. (eds.), Bacteriophage T4. ASM, Washington, D.C. Champness, W.C. and L. Snyder. 1982. The ggi site: A gig- acting bacteriophage T4 regulatory region that can affect expression of all the T4 late genes. J. Mol. Biol. iEE: 395-407. Champness, W.C. and L. Snyder. 1984. Bacteriophage T4 gol site: sequence analysis and effects of the site on plasmid transformation. J. Virol. EQ: 555-562. Chang, A.C.Y. and S.N. Cohen. 1978. Construction and characterization of amplifiable multicopy DNA cloning vehicles derived from the p15A cryptic miniplasmid. J. Bact. igg: 1141-1156. Cooley, W., K. Sirotkin, R. Green, and L. Snyder. 1979. A new gene of Escherichia coli K-12 whose product participates in T4 bacteriophage late gene expression: interaction of ii: with the T4-induced polynucleotide 5'-kinase 3'-phosphatase. J. Bact. igg: 83-91. Daegelen, P., Y.D. D'Aubeuton—Carafa and E. Brody. 1982. The role of rho in bacteriophage T4 development. I. Control of growth and polarity. Virol. 117: 105-120. Hall, M.N., J. Gabay and M. Schwartz. 1983. Evidence for coupling of synthesis and export of an outer membrane protein in Escherichia coli. EMBO J. E: 15-19. Jacobs, K.A., L.M. Albright, D.K. Shibata and E.P. Geiduschek. 1981. Genetic complementation by cloned bacteriophage T4 late genes. J. Virol. E2: 31-45. Jacobs, K.A. and E.P. Geiduschek. 1981. Regulation of expression of cloned bacteriophage T4 late gene 23. J. Virol. E2: 46-59. 83 84 Kao, C., E. Gumbs and L. Snyder. 1987. Cloning and characterization of the Escherichia coli lit gene, which blocks bacteriophage T4 late gene expression. J. Bact. igg: 1232-1238. Karam, J.D., L. Gold, B.S. Singer, and M. Dawson. 1981. Translational regulation: identification of the site on bacteriophage T4 rIIB and mRNA recognized by the regA gene function. P.N.A.S. ZE: 4669-4673. Laemmli, U.K. 1970. Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature 227: 680-685. Laemmli, V.K., F. Beguin and G. Gujer-Kellenberger. 1970. A factor preventing the major head protein of bacteriophage T4 from random aggregation. J. Mol. Biol. 47: 69—85. Maloy, S.R. and W.D. Nunn. 1981. Selection for loss of tetracycline resistance by Escherichia coli. J. Bacteriol. 145: 1110-1112. Maniatis, T., E.F. Fritsch and J. Sambrook. 1982. Molecular Cloning, A Laboratory Manual. Cold Spring Harbor Laboratory, New York. McPheeters, D.S., A. Christensen, E.T. Young, G. Stormo, and L. Gold. 1986. Translational regulation of expression of the bacteriophage T4 lysozyme gene. Nucl. Acids Res. ii: 5813-5826. Messing, J. 1979. A multi-purpose cloning system based on the single-stranded DNA bacteriophage M13. Recombinant DNA Technical Bulletin, NIH Publication No. 79-99, 2, No. 2: 43-48. Miller, J.H. 1972. Experiments in Molecular Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor New York. Parker, M.L., A.C. Christensen, A. Boosman, J. Stockard, E.T. Young, and A.H. Doermann. 1984. Nucleotide sequence of bacteriophage T4 gene 23 and the amino acid sequence of its product. J. Mol. Biol. iEQ: 399-416. Rabussay, D. 1983. Phage-evoked changes in RNA polymerase, p. 167-173. In C.K. Mathews, E.M. Kutter, G. Mosig and P.B. Berget (eds.), Bacteriophage T4. ASM, Washington, D.C. 85 Runnels, J. and L. Snyder. 1978. Isolation of a bacterial host selective for bacteriophage T4 containing cytosine in its DNA. J. Virol. EZ: 815-818. Selzer, G., A. Bolle, H. Krisch and R. Epstein. 1978. Construction and properties of recombinant plasmids containing the rII genes of bacteriophage T4. Mol. Gen. Genet. igg: 301-309. Shapira, S.K., J. Chou, F.V. Richaud and M.J. Casadaban. 1983. New versatile plasmid vectors for expression of hybrid proteins coded by a cloned gene fused to lacZ gene sequences encoding an enzymatically active carboxy- terminal portion of -galactosidase. Gene EE: 71-82. Simon, L.D. 1972. Infection of Escherichia coli by T2 and T4 bacteriophages as seen in the electron microscope: T4 head morphogenesis. P.N.A.S. g2: 907-911. Snyder, L., L. Gold and E. Kutter. 1976. A gene of bacteriophage T4 whose product prevents true late transcription on cytosine—containing T4 DNA. P.N.A.S. 73: 3098-3102. Wood, W.B. 1966. Host specificity of DNA produced by Escherichia coli: bacterial mutations affecting the restriction and modification of DNA. J. Mol. Biol. ig: 118-133. SUMMARY Perhaps the most interesting aspect of this work is the discovery that a short amino-acid sequence in the ggi region can affect the expression of a gene which contains it. This may represent a new type of retroregulation, which is defined as a regulatory mechanism that functions post- transcriptionally from a gig-acting promoter-distal site (Gottesman et al., 1982). We have proposed that the interaction of the ggi region with the iii protein causes a specific termination within the region, but it is not clear whether this occurs at the level of transcription or translation. The construction of ggi- iggE fusions under the control of the inducible igg promoter has made it possible to test this. To examine whether transcription is affected, these fusions can be put into an E. ggii iii(Con) strain carrying the iggiq allele to prevent expression until the inducer IPTG is added. After induction, the amount and size of iggE RNA produced can be monitored by Northern blot analysis. Since the ggi region in these fusions is close to the amino-terminal end of iggE, it would be difficult to detect the small protein fragment which would result if translation terminated in this region. Therefore, it would be desirable 86 87 to make another fusion of the ggi region somewhere toward the carboxyl-terminal end of iggE. In this way, a prematurely terminated protein would be large enough to be immunoprecipitated with anti-B-galactosidase and visualized on an acrylamide gel. Whether an RNA secondary structure in the ggi region is required for termination can be tested directly by site- specific mutagenesis of the region of dyad symmetry. As mentioned in the manuscript, some of the ggi mutations would disrupt the base pairing in the stem of the predicted hairpin structure. By making a specific second mutation which permits base pairing with the nucleotide changed by the ggi mutation, the putative hairpin structure can be restored. If the wild-type phenotype of the ggi region is also restored, this will prove that the RNA secondary structure is essential for the function of the ggi region. A potential problem with this experiment is that the double mutant construction would also have two amino acid changes in the ggi region. A similar experiment can be designed which would change only the putative RNA structure and not the amino acid sequence. Starting with the wild-type sequence from the ggi region, the third base of a valine codon which is within the region of dyad symmetry could be specifically changed to make a silent mutation. If this construction then had the ggi mutant phenotype, it could only be due to a change in the RNA and not the amino acid 88 sequence. In addition to the gig-effect on genes to which it is internal, the ggi region has also been observed to affect the expression of genes which are closely linked (Champness and Snyder, 1982). This effect could be examined more directly in plasmid clones by inserting the ggi region such that it were properly expressed into a gene adjacent to iggE. The effect of the ggi region on expression of a nearby gene could then be determined by measuring B—galactosidase synthesis. The iiggg-effect of the ggi region on the growth of iii(Con) bacteria can also be examined through use of the ol-iggE fusion plasmids. The experiments with lambda clones of the ggi region (see appendix) suggested that the interaction of the ggi region with the iii_protein could rapidly inhibit the growth of the bacteria. Using the more tightly regulated iggE fusion system, it is possible to ask specific questions about the physiological response to the ol-iii interaction. After induction with IPTG, the synthesis of RNA, DNA, and protein can be easily measured. Such experiments have recently been done using the ggi-iggE fusion plasmids (C. Kao, unpublished). It appears that protein synthesis, but not DNA or RNA synthesis, is specifically and severely inhibited in iig(Con) bacteria within minutes after induction of expression of the wild- type. We do not yet have much information about the function 89 of the iii gene product in E. ggii. Neither the overproduction of the iii protein nor the lack of it has any phenotype for the bacteria. As a genetic approach to understanding iii function, it would be useful to identify second-site suppressors of iii(Con) mutations. The wild-type ol-iggE fusion plasmid could be also used for this purpose. The ggi-iii interaction is almost always lethal, since .iii(Con) bacteria with the wild-type ggi-iggE plasmid soon become inviable if grown always in the presence of IPTG. However, the rare survivors which grow with IPTG may have a mutation which suppresses the iii(Con) phenotype. A genetic analysis of these mutants may reveal other host functions which interact with iii or ggi to mediate the response. We know of no precedent for the effects on gene expression, both in gig and in giggg, which are caused by the ol-iii interaction. The iiggg effect is somewhat reminiscent of other plasmid- or prophage-encoded exclusion systems which block the multiplication of different phages. Two examples are the F-plasmid and phage lambda, which block the multiplication of T7 and T4rII mutants respectively. Both F and lambda encode what are thought to be membrane proteins (pifA,B or rexA,B) which mediate the inhibition. However, in these systems membrane permeability and DNA synthesis are also affected, which is not the case when T4 infects lit(Con). It is possible that these mechanisms share 90 common features, especially if they have a common purpose of defense against phages. On the other hand, it is interesting to speculate that the ii: function might play a normal role in regulation of gene expression. The T4 late gene products are involved in assembly of phage structures and must be maintained in correct intracellular ratios for proper assembly. It seems that this would require more subtle and flexible forms of gene regulation in addition to transcriptional activation. Whether the interaction of ii: with the ggi region is deliberate or adventitious, the discovery of the gig-effect on gene expression may have revealed a novel process of retrogregulation, whereby expression of T4 late genes is coordinated with assembly into capsids. Further study of the iii gene and the T4 ggi region should add to our understanding of regulatory mechanisms operating in prokaryotes. APPENDIX APPENDIX The gol Region of Gene 23 Can Inhibit the Expression of E. coli Genes in Trans Plasmid clones of the wild-type ggi region do not give rise to antibiotic-resistant transformants of iig(Con) bacteria. Among the possible explanations for this phenomenon are that the ggi site affects the uptake, replication or segregation of the plasmid; the gol-containing DNA is degraded in iig(Con) bacteria; or the presence of the ggi site somehow affects the growth of the bacteria. We had indirect evidence to support the latter from transformation experiments with "leaky" plasmid subclones of the wild-type ggi region. These plasmids yielded small, slow-growing antibiotic-resistant transformants of a iig(Con) strain, and when transferred to non-selective media the colonies remained small and slow-growing. This suggested that the plasmid DNA gets into the cells and inhibits growth of the bacteria, since a defect in plasmid replication or segregation could not explain why the bacteria grow slowly on media without antibiotics. To examine more directly the effects of the ggi region on iii(Con) bacteria, T4 DNA containing the ggi region was cloned into a bacteriophage lambda vector. The advantage of this system over plasmid transformation is that phage 91 92 infection at high multiplicity ensures that all of the bacteria in a culture receive the cloned DNA, while not all of the cells in a transformation experiment take up the plasmid DNA. 1.1kb HindIII fragments containing the T4 ggi region, either wild-type or with the ggi2APl3 mutation, were cloned into the unique HindIII site of the vector Charon 7 (Blattner et al., 1977). The cloning site in this vector is in the cI repressor gene, so recombinant phage with DNA inserts can be identified by formation of large clear plaques on E. ggii C600giig. Preparation of lambda DNA, ligations and transfections were performed according to the methods of Davis et al. (1980). Two clones were identified which carried the wild-type T4 DNA or the ggi mutant DNA in the same orientation (A1.1 and al.12AP respectively). When lysates of these clones were spotted on a lawn of E. ggii BKL155 [iii(Con)] at 30°, 1 1.12AP cleared the lawn but A1.1 did not. Both clones gave a clear zone of lysis on BK0155, which lacks the iii gene. Thus, the phenotypes of these lambda clones for growth on iii(Con) Ei ggii were analogous to the transformation phenotypes of the plasmid clones. The lambda clones were then used to infect cultures of BKL155 and BK0155, and the growth of the bacteria was monitored after infection. The bacteria were first grown to 4x108/ml in tryptone broth + 0.2% maltose (TBM), then spun 93 and resuspended in one-tenth volume of 20mM MgSO4. The cultures were each divided into three tubes, to which were added saline (mock infection) or either lambda clone at a multiplicity of ten. After a 10 min. incubation at 30° for phage adsorption, the infected cultures were diluted 1:35 into TBM and incubated at 30° with aeration. The optical density of the cultures was measured at 20 min intervals and the results are shown in Figure 1A. The growth of BKL155 infected with A1.1 (wild-type ggi clone) is strikingly different from the other cultures. Growth of BKL155/Al.l appears to stop prematurely and level off, instead of continuing to increase in mass until lysis begins at about 80 min. post-infection. In fact, the culture never lysed and no phage were produced, as indicated by titer of the culture after 140 min. (data not shown). In contrast, 11.1 showed a normal pattern of infection and phage production in BK0155, as did 11.12AP in both strains. These observations support the conclusion that in a iii(Con) host, the ggi region can act in iiggg to inhibit the growth of the bacteria. Within a very short time after the DNA with the wild-type ggi region enters the cell (less than an hour), the rate of growth is dramatically decreased and growth eventually ceases. This is sufficient to explain the lack of iii(Con) transformants with a wild-type ggi plasmid clone, although effects of the site on plasmid gene expression and replication are not ruled out. '1 Fig. 1A. 94 Growth of E. coli infected with a lambda clone of the T4 gol region. The growth of BKL155 [iii(Con)] and BK0155 (lit°) was followed after infection with X1.1 (gol') or 11.12AP (ggi2AP13) by measuring optical density of the cultures. Symbols: BKL155,- (uninfected),A ,o ; BK0155,0 (uninfected), A,O; 11.1,. ,o ; Al.l2AP,A ,A . 1.0 ., 0.0.625 Fig.1A 1 20 94a l l J J 40 60 80 100 minutes after infection B IBL IOGRAPHY Bibliography Artman, M. and S. Werthamer. 1974. Transition from streptomycin sensitive to streptomycin-resistant protein synthesis during bacteriophage T4 development. Biochem. Biophys. Res. Commun. 22: 75-81. Benzer, S. 1959. On the topology of genetic fine structure. P.N.A.S. 3E: 1607-1620. Binkowski, G. and L.D. Simon. 1983. Host functions that affect T4 reproduction, p. 342-350.‘ In C.K. Mathews, E.M. Kutter, G. Mosig, and P.B. Berget (eds.). Bacteriophage T4. A.S.M., Washington, D.C. Black, L.W. and M.K. Showe. 1983. Morphogenesis of the T4 head, p. 219—245. In C.K. Mathews, E.M. Kutter, G. Mosig, and P.B. Berget (eds.) Bacteriophage T4 A.S.M., Washington, D.C. Blattner, F.R., B.G. Williams, A.E. Blechl, K. Denniston- Thompson, H.E. Faber, L.A. Furlong, D.J. Grunwald, D.O. Kiefer, D.D. Moore, J.W. Schumm, E.L. Sheldon, and O. Smithies. 1977. Charon phages: safer derivatives of bacteriophage lambda for DNA cloning. Science i2E: 161-169. Bolle, A., R.H. Epstein, W. Salser, and E.P. Geiduschek. 1968. Transcription during bacteriophage T4 development: requirements for late messenger synthesis. J. Mol. Biol. 22: 339-362. Brenner, S., F. Jacob and M. Meselson. 1961. An unstable intermediate carrying information from genes to ribosomes for protein synthesis. Nature 190: 576-581. Brody, E.N., H. Diggelmann, and Gieduschek. 1970. Transcription of the bacteriophage T4 template. Obligate synthesis of T4 prereplicative RNA in vitro. Biochemistry 2: 1300-1309. Brody, E., D. Rabussay, and D.H. Hall. 1983. Regulation of transcription of prereplicative genes, p. 174—183. In C.K. Mathews, E.M. Kutter, G. Mosig, and P. Berget (eds.), Bacteriophage T4. A.S.M., Washington, D.C. Brody, E., R. Sederoff, A. Bolle, and R.H. Epstein. 1970. Early transcription in T4-infected cells. Cold Spring Harbor Symp. Quant. Biol. 2E: 203-211. 95 96 Cardillo, T.S., E.F. Landry, and J.S. Wiberg. 1979. regA protein of bacteriophage T4D: Identification, schedule of synthesis and autogenous regulation. J. Virol. 22: 905-916. Caruso, M., A. Coppo, A. Manzi, and J. Pulitzer. 1979. Host-virus interactions in the control of T4 prereplicative transcription I. tabC (rho) mutants. J. Mol. Biol. i22: 950-977. Champness, W.C. and L. Snyder. 1982. The ggi site: A gig— acting bacteriophage T4 regulatory region that can affect expression of all the T4 late genes. J. Mol. Biol. i22: 395-407. Champness, W.C. and L. Snyder. 1984. Bacteriophage T4 gol site: sequence analysis and effects of the site on plasmid transformation. J. Virol. 22: 555-562. Christensen, A. C. and E.T. Young. 1982. T4 late transcripts are initiated near a conserved DNA sequence. Nature 299: 369-371. Christensen, A.C. and E.T. Young. 1983. Characterization of T4 transcripts, pp. 184-88. In C. Mathews, E.M. Kutter, G. Mosig, and P.B. Berget (eds.). Bacteriophage T4, A.S.M., Washington, D.C. Chu, F.K., G.F. Maley, M. Belfort and F. Maley. 1985. ig vitro expression of the Intron-containing gene for T4 phage thymidylate synthase. J. Biol. Chem. 260: 10680- 10688. Chu, F.K., G.F. Maley, F. Maley, and M. Belfort. 1984. Intervening sequence in the thymidylate synthese gene of bacteriophage T4. P.N.A.S. 2i: 3049-3053. Chu, F.K., G.F. Maley, D.K. West, M. Belfort, and F. Maley. 1986. Characterization of the intron in the phage T4 thymidylate synthase gene and evidence for its self excision from the primary transcript. Cell 22: 157- 166. Cooley, W., K. Sirotkin, R. Green, and L. Snyder. 1979. A new gene of Escherichia coli K-12 whose product participates in T4 bacteriophage late gene expression: interaction of iii with the T4-induced polynucleotide 5'-kinase 3'-phosphatase. J. Bact. igg: 83-91. 97 Coppo, A., A. Manzi, J.F. Pulitzer, and H. Takahashi. 1975. Host mutant (tabD)-induced inhibition of bacteriophage T4 late transcription. 1. Isolation and phenotypic characterization of the mutants. J. Mol. Biol. 22: 579- 600. Crick, F.H.C., L. Barnett, S. Brenner, and R.J. Watts-Tobin. 1961. General nature of the genetic code for proteins. Nature 192: 1227-1232. Daegelen, P., Y. D'Aubeuton-Carafa, and E. Brody. 1982. The role of rho in bacteriophage T4 development. I. Control of growth and polarity. Virology 117: 105-120. Davis, R.W., D. Botstein, and J.R. Roth. 1980. Advanced Bacterial Genetics. Cold Spring Harbor Laboratory, Cold Spring Harbor, New York. de Franciscis, V. and E. Brody. 1982. ig vitro system for middle T4 RNA I. Studies with Escherichia coli RNA Polymerase. J. Biol. Chem. 257: 4087-4096. Delbruck, M. and W.T. Bailey. 1946. Induced mutations in bacterial viruses. Cold Spring Harbor Symp. Quant. Biol. ii: 33-37. Doermann, A.H., F.A. Eiserling, and L. Boehner. 1973. Genetic control of capsid length in bacteriophage T4. I. Isolation and preliminary description of four new mutants. J. Virol. i2: 374-385. Doherty, D.H. 1982. Genetic studies on capsid-length determination in bacteriophage T4. I. Isolation and partial characterization of second-site revertants of a gene 23 mutation affecting capsid length. J. Virol. 22: 641-654. Dube, S.K. and P.S. Rudland. 1970. Control of translation by T4 phage: altered binding of disfavoured messengers. Nature 226: 820-823. Eiserling, F.A., J. Corso, S. Feng, and R.H. Epstein. 1984. Intracellular morphogenesis of bacteriophage T4 II. Head morphogenesis. Virology 137: 95-101. Eiserling, F.A., E.P. Geiduschek, R.H. Epstein, and E.J. Matter. 1970. Capsid size and deoxyribonucleic acid length: the petite variant of bacteriophage T4. J. Virol. 2: 865—876. 98 Elliot, T.E. and E.P. Geiduschek. 1984. Defining a bacteriophage T4 late promoter: absence of a "-35" region. Cell 22: 211-219. Elliot, T.E., E.P. Geiduschek, and G.A. Kassavetis. 1984. The complex pattern of transcription in the segment of the bacteriophage T4 genome containing three of the head protein genes. Virology i22: 260-282. Geiduschek, E.P., T. Elliott, and G.A. Kassavetis. 1983. Regulation of late gene expression, pp. 189-192. In C.K. Mathews, E. Kutter, G. Mosig, and P.B. Berget (eds.) Bacterigphgge T4, A.S.M., Washington, D.C. Geiduschek, E.P. and O. Grau. 1970. T4 anti-messenger, pp. 190—203. In L. Silvestri (ed.), RNA polymerase and transcription. North-Holland Publishing Co., Amsterdam. Georgopoulos, C.P., R.W. Hendrix, A.D. Kaiser, and W.B. Wood. 1972. Role of the host cell in bacteriophage morphogenesis: effects of a bacterial mutation on T4 head assembly. Nature New Biol. 222: 38-41. Goff, C.G. 1974. Chemical structure of modification of the Escherichia coli ribonucleic acid polymerase alpha polypeptides induced by bacteriophage T4 infection. J. Biol. Chem. 222: 6181-6190. Gold, L., P. O'Farrell, and M. Russel. 1976. Regulation of gene 32 expression during bacteriophage T4 infection of Escherichia coli. J. Biol. Chem. 251: 7251-7262. Gott, J.M., D.A. Shub, and M. Belfort. 1986. Multiple self- splicing introns in bacteriophage T4: evidence from autocatalytic GTP labeling of RNA ip vitro. Cell 22: 81-87. Gottesman, M., A. Oppenheim and D. Court. 1982. Retroregulation: control of gene expression from sites distal to the gene. Cell 22: 727-728. Greener, A. and C. Hill. 1980. Identification of a novel genetic element in Escherichia coli K-12. J. Bact. 144: 312-321. Guha, A., W. Szybalski, w. Salser, A. Bolle, E.P. Geiduschek and J.F. Pulitzer. 1971. Control and polarity of transcription during bacteriophage T4 development. J. Mol. Biol. 22: 329-349. 99 Hattman, S. and T. Fukasawa. 1963. Host-induced modification of the T-even phages due to defective glucosylation of their DNA. P.N.A.S. 22: 297-300. Hattman, S. and P.H. Hofschneider. 1968. Interference of bacteriophage T4 in the reproduction of RNA phage M12. J. Mol. Biol. 22: 173-190. Helland, D.E. 1979. Complementary transcribed T4 RNA. Association with the polysomes. Biochem. Biophys. Res. Comm. 2i: 91-98. Hershey, A.D. and M. Chase. 1952. Independent functions of viral protein and nucleic acid in growth of bacteriophage. J. Gen. Physiol. 22: 39—56. Hershey, A.D. and R. Rotman. 1948. Linkage among genes controlling inhibition of lysis in a bacterial virus. P.N.A.S. 22: 89-96. Horvitz, H.R. 1973. Polypeptide bound to the host RNA polymerase is specified by a T4 control gene 33. Nature New. Biol. 244: 137-140. Hsu, W.T. and S.B. Weiss. 1969. Selective Translation of T4 template RNA by ribosomes from T4-infected Escherichia coli. P.N.A.S. 22: 345-351. Jabbar, M.A. and L. Snyder. 1984. Genetic and physiological studies of an Escherichia coli locus that restricts polynucleotide kinase-and RNA ligase-deficient mutants of bacteriophage T4. J. Virol. 2i; 522-529. Jacobs, K.A., L.M. Albright, D.K. Shibata, and E.P. Geiduschek. 1981. Genetic complementation by cloned bacteriophage T4 late genes. J. Virol. 22: 31—45. Jacobs, K.A. and E.P. Geiduschek. 1981. Regulation of expression of cloned bacteriophage T4 late gene 23. J. Virol. 22: 46-59. Kao, C., E. Gumbs, and L. Snyder. 1987. Cloning and characterization of the Escherichia coli lit gene, which blocks bacteriophage T4 late gene expression. J. Bact. i22: 1232-1238. Karam, J.D., L. Gold, B.S. Singer, and M. Dawson. 1981. Translational regulation: identification of the site on bacteriophage T4 rIIB and mRNA recognized by the regA gene function. P.N.A.S. 22: 4669-4673. 100 Kassavetis, G.A., T. Elliott, and E.P. Geiduschek. 1984. Regulation of transcription in the segment of the bacteriophage T4 genome containing three of the head protein genes: plasmid-specific and phage chromosome- specific effects. Virology i22: 283-294. Kassavetis, G.A., T. Elliot, D.P. Rabussay, and E.P. Geiduschek. 1983. Initiation of transcription at phage T4 late promoters with purified RNA polymerase. Cell 22: 887-897. Kassavetis, G.A. and E.P. Geiduschek. 1982. Bacteriophage T4 late promoters: mapping 5' ends of T4 gene 23 mRNAs. EMBO J. i: 107-114. Kassavetis, G. and E.P. Geiduschek. 1984. Defining a bacteriophage T4 late promoter: bacteriophage T4 gene 55 protein suffices for directing late promoter recognition. P.N.A.S. 2i: 5101-5105. Kennell, D. 1970. Inhibition of host protein synthesis during infection of Escherichia coli by bacteriophage T4. II. Induction of host messenger ribonucleic acid and its exclusion from polysomes. J. Virol. 9: 208-217. Knight, J.A., L.W. Hardy, D. Rennell, D. Herrick, and A.R. Poteete. 1987. Mutations in an upstream regulatory sequence that increase expression of the bacteriophage T4 lysozyme gene. J. Bact. i22: 4630-4636. Komeda, Y. 1986. Transcriptional control of flagellar genes in Escherichia coli K-12. J. Bact. 168: 1315-1318. Kutter, E., A. Beug, R. Sluss, L. Jensen and D. Bradley. 1975. The production of undergraded cytosine-containing DNA by bacteriophage T4 in the absence of CTPase and endonucleases II and IV, and its effects on T4-directed protein synthesis. J. Mol. Biol. 22: 591-607. Laemmli, U.K., F. Beguin, and G. Gujer-Kellenberger. 1970. A factor preventing the major head protein of bacteriophage T4 from random aggregation. J. Mol. Biol. 47: 69—85. Luria, S. 1947. Reactivation of irradiated bacteriophage by transfer of self-reproducting units. P.N.A.S. 22: 253- 264. Luria, S.E. and M.L. Human. 1952. Nonhereditary, host- induced variation of bacterial viruses. J. Bacteriol. 22: 557-569. 101 Malik, S., M. Dimitrov, and A. Goldfarb. 1985. Initiation of transcription by bacteriophage T4—modified RNA polymerase independently of host sigma factor. J. Mol. Biol. i22: 83-91. Malik, S. and A. Goldfarb. 1984. The effect of a bacteriophage T4-induced polypeptide on host RNA polymerase interaction with promoters. J. Biol. Chem. 222: 13292—13297. Mattson, T., J. Richardson, and D. Goodin. 1974. Mutant of bacteriophage T4D affecting expression of many early genes. Nature 250: 48—50. Mattson, T., G. van Houwe, A. Bolle, G. Selzer, and R. Epstein. 1977. Genetic identification of cloned fragments of bacteriophage T4 DNA and complementation by some clones containing early T4 genes. Mol. Gen. Genet. i22: 318-326. McPheeters, D.S., A. Christensen, E.T. Young, G. Stormo, and L. Gold. 1986. Translational regulation of expression of the bacteriophage T4 lysozyme gene. Nucl. Acids. Res. i2: 5813-5826. Miller, E.S., J. Karam, M. Dawson, M. Trojanowska, P. Gauss, and L. Gold. 1987. Translational repression: biological activity of plasmid-encoded bacteriophage T4 regA protein. J. Mol. Biol. i22: 397-410. Mosig, G. 1983. T4 genes and gene products. pp. 362-374. In. C.K. Mathews E.M. Kutter, G. Mosig and P.B. Berget (eds.), Bacteriophage T4. A.S.M., Washington, D.C. Nomura, M., R. Gourse, and G. Baughman. 1984. Regulation of the synthesis of ribosomes and ribosomal components. Ann. Rev. Biochem. 22: 75-117. O'Farrell, P.Z. and L.M. Gold. 1973. Bacteriophage T4 gene expression: evidence for two classes of prereplicative cistrons. J. Biol. Chem. 248: 5502-5511. Pearson, R.E. and L. Snyder. 1980. Shutoff of lambda gene expression by bacteriophage T4: role of the T4 alc gene. J. Virol. 22: 194-202. Pulitzer, J.F., M. Colombo, and M. Ciaramella. 1985. New control elements of bacteriophage T4 pre-replicative transcription. J. Mol. Biol. 182: 249—263. 102 Pulitzer, J.F., A. Coppo, and M. Caruso. 1979. Host-virus interactions in the control of T4 prereplicative transcription II. Interaction between tabC (rho) mutants and T4 ggi mutants. J. Mol. Biol. i22: 979-997. Rabussay, D. 1983. Phage-evoked changes in RNA polymerase, p. 167-173. In C.K. Mathews, E.M. Kutter, G. Mosig, and P. Berget (eds.), Bacteriophage T4. A.S.M., Washington, D.C. Rabussay, D. and E.P. Geiduschek. 1977b. Phage T4-modified RNA polymerase transcribes T4 late genes ip vitro. P.N.A.S. 23: 5305—5309. Rabussay, D. and E.P. Geiduschek. 1977a. Regulation of gene action in the development of lytic bacteriophages. Compr. Virol. 2: 1—196. Ratner, D. 1974a. Bacteriophage T4 transcriptional control gene 55 codes for a protein bound to E. coli RNA polymerase. J. Mol. Biol. 22: 803—807. Ratner, D. 1974b. The interactions of bacterial and phage proteins with immobilized E. coli RNA polymerase. J. Mol. Biol. 22: 373-383. Revel, H.R. 1967. Restriction of non-glycosylated T—even bacteriophage: properties of permissive mutants of E coli B and K12. Virology 2i: 688—701. Riva, S., A. Cascino, and E.P. Geiduschek. 1970a. Coupling of late transcription to viral replication in bacteriophage T4 development. J. Mol. Biol. 22: 85—102. Riva, S., A. Cascino, and E.P. Geiduschek. 1970b. Uncoupling of late transcription from DNA replication in bacteriophage development. J. Mol. Biol. 22: 103—119. Rosenberg, M. and D. Court. 1979. Regulatory sequences involved in the promotion and termination of RNA transcription. Ann. Rev. Genet. i2: 319-353. Runnels, J. and L. Snyder. 1978. Isolation of a bacterial host selective for bacteriophage T4 containing cytosine in its DNA. J. Virol. 22: 815-818. Salser, W., A. Bolle and R. Epstein. 1970. Transcription during bacteriophage T4 development: a demonstration that distinct subclasses of "early" RNA appear at different times and that some are "turned off” at late times. J. Mol. Biol. 22: 271—295. 103 Sarabhai, A.S., A.C.W. Stretton and S. Brenner. 1964. C0- linearity of the gene with the polypeptide chain. Nature 201: 13-17. Shinedling, S., M. Gayle, D. Pribnow, and L. Gold. 1987. Mutations affecting translation of the bacteriophage T4rIIB gene cloned in Escherichia coli. Mol. Gen. Genet. 222: 224-232. Simon, L.D. 1972. Infection of Escherichia coli by T2 and T4 bacteriophages as seen in the electron microscope: T4 head morphogenesis. P.N.A.S. 22: 907-911. Simon, L.D., T.J.M. McLaughlin, D. Snover, J. Ou, C. Grisham, and M. Loeb. 1975. E. coli membrane lipid alteration affecting T4 capsid morphogenesis. Nature 256: 379-383. Sirotkin, K., W. Cooley, J. Runnels, and L. Snyder. 1978. A role in true-late gene expression for the T4 bacteriophage 5' polynucleotide kinase 3' phosphatase. J. Mol. Biol. i22: 221-233. Sirotkin, K., J. Wei, and L. Snyder. 1977. T4 bacteriophage-coded RNA polymerase subunit blocks host transcription and unfolds the host chromosome. Nature 222: 28-32. Sjoberg, B.M., S. Hahne, C.Z. Mathews, C.K. Mathews, K.N. Rand, and M.J. Gait. 1986. The bacteriophage T4 gene for the small subunit of ribonucleotide reductase contains an intron. EMBO J. 2: 2031-2056. Smith, F.L. and R. Haselkorn. 1969. Proteins associated with ribosomes in T4-infected E. coli. Cold Spring Harbor Symp. Quant. Biol. 22: 91—94. Snyder, L., L. Gold and E. Kutter. 1976. A gene of bacteriophage T4 whose product prevents true late transcription on cytosine-containing T4 DNA. P.N.A.S. 22: 3098-3102. Snustad, D.P., N. Haas, D.G. Oppenheimer. 1986. Bacteriophage T4 regulatory proteins gp ggi/gig binds to DNA in the absence of RNA polymerase. J. Virol. 22: 1145-1147. 104 Snustad, D.P., M.A. Tigges, K.A. Parson, C.J.H. Bursch, F.M. Caron, J.F. Koerner, and D.J. Tutas. 1976. Identification and preliminary characterization of a mutant defective in the bacteriophage T4-induced unfolding of the Escherichia coli nucleoid. J. Virol. 17: 622-641. Stevens, A. 1972. New small polypeptides associated with DNA-dependent RNA polymerase of Escherichia coli after infection with bacteriophage T4. P.N.A.S. 69: 603—607. Stevens, A. 1977. Inhibition of DNA—enzyme binding by an RNA polymerase inhibitor from T4 phage-infected Escherichia coli. Biochem. Biophys. Acta 475: 193—196. Streisinger, G., R.S. Edgar, and G.H. Denhardt. 1964. Chromosome structure in phage T4. I. Circularity of the linkage map. P.N.A.S. 2i: 775—779. Thermes, C. and E. Brody. 1984. T4—induced antipolarity: temporal heterogeneity in response of early transcription units. J. Virol. 22: 191—201. Uzan, M., J. Leauty, Y. d'Aubenton-Carafa, and E. Brody. 1983. Identification and biosynthesis of the bacteriophage T4 mot regulatory protein. EMBO J. 2: 1207—1212. Volker, T.A., B. Keller, and T.A. Bickle. 1984. Deletion analysis of a bacteriophage T4 late promoter. Gene 22: 207—213. Wiberg, J.S. and J.D. Karam. 1983. Translational regulation in T4 phage development, p. 193—201. In C.K. Mathews, E.M. Kutter, G. Mosig, and P.B. Berget (eds.). Bacteriophage T4. A.S.M., Washington, D.C. Williams, K.P., G.A. Kassavetis, F.S. Esch, and E.P. Geiduschek. 1987. Identification of the gene encoding an RNA polymerase—binding protein of bacteriophage T4. J. Virol. 2i: 597-599. Wood, W.B. and H.R. Revel. 1976. The Genome of bacteriophage T4. Bacteriol. Rev. 22: 847-868. Wu, R., E.P. Geiduschek and A. Cascino. 1975. The role of replication proteins in the regulation of bacteriophage T4 transcription. II. Gene 45 and late transcription uncoupled from replication. J. Mol. Biol. 96: 539-562. 105 Young, E.T., T. Mattson, G. Selzer, G. van Houwe, A. Bolle, and R. Epstein. 1980. Bacteriophage T4 gene transcription studied by hybridization to cloned restriction fragments. J. Mol. Biol. i22: 423—445. Young, E.T., R.C. Menard, and J. Harada. 1981. Monocistronic and polycistronic bacteriophage T4 gene 23 messages. J. Virol. 22: 790-799. "IIIIIIIIIIIIIIIES