Studies on bacteriophage and host genes involved in morphogenesis.

Studies on characteristics of Staphylococcus aureus the staphylococcal alpha hemolysin or exotoxin in cell free filtrates, rather than coagulase, correlated with the lethal effects in suckling mice. As observed in these studies alpha hemolysin production was increased by prolonged incubation of S. aureus cultures in the presence of approximately 25% increased carbon dioxide. In normal atmosphere some strains of S. aureaus tested failed to produce alpha hemolysin. Staphylococcal strain differences were observed in hemolysin tube test with human red blood cells. Some S. aureaus strains showed “hot†and “cold†lysis, some only “cold†lysis and some failed to lyse human cells. Staphylococci formed 2 hemolytic bands around S. aureaus colonies on rabbit blood agar after incubation at 37°C for 24 fours. Two to 3 additional bands were formed after refrigeration at 4°C for 24 hours. S. aureaus colonies produced only 1 hemolytic zone on human blood agar. Tube and plate hemolysin test using rabbit red blood cells showed “hot†and “cold†hemolysin activity. Alpha hemolysin was believed to be responsible for “hot†and “cold†hemolysis in tube tests. This “hot-cold†reaction by alpha hemolysin may be related to the “hot†and “cold†hemolytic bands formed around S. aureus colonies when cultured on rabbit blood agar. No immunological difference was established between “hot†and “cold†alpha hemolysins. The presence of “cold†hemolysin in blood agar however does not rule out other unidentified hemolysins. Ultraviolet irradiation and boiling inactivated lethal toxin and hemolysin in a straight line relationship. The “cold†hemolysin however was more resistant to these physical agents than the “hot†hemolysin. Necrosis was produced in sucking mice inoculated with material containing “cold†hemolysin. The necrotic activity correlated the cold hemolysin and is believed to be a new finding. Concentrated alpha hemolysin showed an equivalent 1:1600 minimal hemolytic dose as measure by rabbit red blood cell lysis and an equivalent of 1:2000 minimal lethal dose as measured by intramuscular inoculation of sucking mice. The virulence of S. aureus strain 10 appeared to be related to the number of organisms’ present rather than alpha toxin. Although the virulence of staphylococcus appeared to be related to number of organisms injected into sucking mice, virulence of the other strains tested was associated more specifically with the formation of alpha hemolysin. Other virulence mechanisms should be considered in S. aureaus strains such as strain 10 unless such strains produce toxins in-vivo. Some protection of rabbits was shown when they were inoculated with S. aureus strain 15 mixed and incubated with rabbit or human plasma prior to inoculation. Results were variable in rabbit protection tests which may be explained on differences in the toxic content of cultures due to bacterial variation or due to variation in the different groups of rabbits selected for the test or a combination of both.

STUDIES ON BACTERIOPHAGE AND HOST GENES INVOLVED IN MORPHOGENESIS by Kathryn Louise Tilly A dissertation submitted to the faculty of The University of Utah in partial fulfillment of the requirementsfor the degree of Doctor of Philosophy Department of Cellular, Viral and Molecular Biology The University of Utah August 1982 © 1982 Kathryn Louise Tilly All Rights Reserved THE UNIVERSITY OF UTAH GRADUATE SCHOOL SUPERVISORY COMMITTEE APPROVAL of a dissertation submitted by Kathryn Louise Tilly This dissertation has been read by each member of the following supervisory· committee and by majority vote has been found to be satisfactory. July 7, 1982 Chainnan: Cos ta Georgopoulos t!t~t Iff2 fJU7< 1982 ~982 THE UNIVERSITY OF UTAH OF GRADUATE SCHOOL FINAL READING APPROVAL To the Graduate Council of The University of Utah: I have read the dissertation of Ka thryn Lou; se Ti 11 Y in its final form and have found that (1) its format, citations, and bibliographic style are consistent and acceptable; (2) its illustrative materials including figures, tables, and charts are in place; and (3) the final manuscript is satisfactory to the Supervis­ory Committee and is ready for submission to the Graduate School. July 7, 1982 Date Costa Georgopoulos Chairperson, Supervisory Committee Chairman I Dean Approved for the Graduate Council James L. Cl ayto Dean of The Grad uate School ABSTRACT It was shown previously that the Escherichia coli groE locus is required for bacterial growth, the head assembly of bacteri ophages A and T4, and bacteri ophage T5 tai 1 assemb ly. In thi s dissertation, it is shown by genetic and biochemical analysis of AgroE transducing phages that two closely linked genes, groEL and groES make up the groE locus. So far, only groEL- mutants block T4 morphogenesis. The product of the groES gene is a protein of 15,000 Mr , while the groEL gene encodes the previously identified groE gene product, a protein of 65,000 Mr. The products of the two groE genes interact functionally, since it was shown that mutations in the groEL gene suppress groES- mutations. Finally, it is shown that both of the groE gene products belong to the "heat-shock" group of I. coli proteins, regulated by the htpR locus. These two genes appear to be regulated at the transcriptional level because they are both transcribed from a single promoter whose activity is modified by heat shock. ABSTRACT • • • • ACKNOWLEDGEMENTS TABLE OF CONTENTS Chapter ••••• · . . I. BACTERIOPHAGE-HOST INTERACTIONS IN MORPHOGENESIS. • • • • • • • . . . . . . . . . Background • • • • • • • • • • • • • • • • • • • • • Function of the groE Genes in Bacterial Assembly Effect of groE Mutations on Bacterial Growth . . Other Host Genes Involved in Phage Assembly. Literature Cited •••••••••••••••••• II. IDENTIFICATION OF A SECOND ESCHERICHIA COLI GROE GENE WHOSE PRODUCT IS NECESSARY FO-R--­BACTERIOPHAGE MORPHOGENESIS ••• Introduction ••••••••• · . . . . . Methods and Materials. . . . . . . . Results ••••••••• Discussion. • • • • • • ••••• Literature Cited •••••••• · . . III. EVIDENCE THAT THE TWO ESCHERICHIA COLI GROE MORPHOGENETIC GENE PRODUCTS INTERACT IN VIVO •• Materials and Methods. . . . . . . Results ••••• Discussion •••• Acknowledgements •• Literature Cited •• Appendi x • • • • . .. .. .. .. ... ... .. . . .. ... ... .. .. IV. GROE GENE REGULATION • Introduction •••••• . . . . . . . . . Page iv vii 1 1 4 18 24 27 31 31 32 37 51 56 58 58 59 61 63 63 65 70 70 Methods and Materials. Results •••••••• Discussion •••••• Literature Cited ••••• v. DISCUSSION AND FUTURE DIRECTIONS. vi Page 72 74 90 94 96 ACKNOWLEDGEMENTS I would like to thank Costa Georgopoulos for his endless patience and guidance throughout the course of these studies. I would also like to acknowledge my other Advisory Committee members, Sherwood Casjens, Ellie Ehrenfeld, Ray Gesteland and Donald Summers, for their advice and constructive criticism. Glenn Herrick, Dean Dawson, Sam Cartinhour, Kevin Anderson and Steve Rice provided expert advice in the molecular biological experiments. Niki McKittrick, Tim Hey and Lyn McDivitt, helped with some of the experiments. This work was supported by Public Health Service Grants GM23917 and GM07464 from the National Institutes of Health. Chapter II is an exact copy of a paper published in the Proceedings of the National Academy of Sciences USA (~:1629-1633, 1981). CHAPTER I BACTERIOPHAGE-HOST INTERACTIONS IN MORPHOGENESIS Background Bacteriophages and their hosts have evolved together so their interactions are many and diverse. Phages use much of the cell's machinery in their life cycles and they, in turn, may benefit their hosts (1). Some viruses, e.g. bacteriophage T4, supply their own DNA polymerase and accessory proteins but many smaller phages (e.g. $ X174) are constra.ined by their genome sizes so they use mostly host enzymes for replication (2). Even the small phages encode proteins that bias the host polymerase toward their replication late in the infectious cycle. Most phages also use the bacterial DNA­directed RNA polymerase for transcription but many (e.g., T4 and TS) modify that enzyme so that it recognizes different sets of promoters (3). Even more independent is phage N4, which encodes its own virion­associated RNA polymerase (4). All phages use the host translation machi nery. There also appear to be interactions on more subtle levels. One way of searching the host genome for other genes whose products are required for viral growth is to isolate host mutants in which the virus is unable to multiply or function normally. Such an approach has been particularly fruitful in the bacteriophage A-I. coli genetic system. The expected mutants to which A cannot adsorb or into which A cannot inject its DNA were isolated (5,6). Many unexpected interactions involving new host genes were also discovered. Bacteria with mutations affecting transcription (7), DNA replication (8), integration (9), and morphogenesis (10) have been isolated. This dissertation will describe the analysis of the I. coli groE genes, whose products are required for both the correct assembly of several phages and for bacterial growth. I. coli groE- mutants were isolated in a general screen for mutants blocking productive growth (10). Cells mutagenized with N- methyl-N'-nitro-N-nitrosoguanidine were plated in the presence of low levels of A and 434 phage. Under these conditions, mutants blocking phage development form large colonies because they do not amplify the phage on the plate (8). Upon further analysis (see below; 10), phage infecting groE- mutants were shown to replicate and transcribe their DNA normally. The normal phage protein profiles were found and the hosts were lysed on time. The phage yields, however, were shown to be reduced to less than one infectious phage per bacterium. Subsequent in vitro complementation, electron microscopic, and protein analysis showed that although functional tails were formed in groE- bacteria, the only head-related structures present were aberrant (e.g., tubes, polyheads, and prohead-like forms with abnormal protein composition). When plating properties of other phage were tested in groE­hosts, it was found that the growth of all lambdoid phages was 2 blocked by all mutants. Phages 186, T4, and T5 were unable to grow in various subsets of the mutant collection. T1, T3, P1, T7, and P2 were able to grow well on all groE- strains. All groE mutations have been mapped to 93 minutes on the revised I coli genetic map (12). Fine structure mapping (13,14) has placed the mutations between the cadA and ampC genes. Subsequent analysis (see below) has shown that there are two ~ genes (denoted groEL and groES), both of which map at 93 minutes. A groE+ transducing phages (15-17) were used to identify the product of the groEL gene (15,16). Georgopoulos and Hohn (15) identified a 65,OOO-Mr protein labeled after infection of cells irradiated with ultra-violet light (UV) with A groE+ transducing phage. The protein's mobility upon NaDodS04-polyacrylamide gel electrophoresis was altered in mutant phage that were unable to grow on a subset of groE- mutants (now identified as the groEL-mutants). Hendrix and Tsui (16) identified the same protein after infection with groE+ phage. They isolated transducing phage mutants with amber mutations in the groEL gene, which formed plaques on I. col i groE+ or groEL - ~, but not groEL -~+ bacteri a. These phage programmed the synthesis of the 65,OOO-Mr protein in UV-irradiated ~ and supE bacteria, but not in sup+ hosts, confirming that this protein was the groEL gene product. The construction of AgroE+ transducing phages simplified the purification of large amounts of groEL gene product. Lysates were made by inducing lysogens of cI857 derivatives of the AgroE+ transducing phage. These were fractionated by standard biochemical 3 methods to obtain nearly homogeneous preparations of gpgroEL (18,19). Electron microscopic analysis of the purified protein showed that it forms a doughnut-like structure (12.6 nm in diameter), which is most probably a double ring with seven-fold rotational symmetry. This is consistent with sedimentation studies, which show the native oligomer to sediment at 255, corresponding to 11-15 sUbunits of 65,000-Mr. A protein found in partially pure RNA polymerase preparations, looks identical to gpgroEL in the electron microscope (20-22). In fact, Hendrix (18) found that purified gpgroEL has an ATPase activity and sedimentation coefficient similar to those of the RNA polymerase contaminant (23,24) so the two proteins are probably identical. In the remaining sections of Chapter I, I will describe in detail (a) the processes affected by groE- mutations and the stages at which they are blocked, (b) what is known about the regulation of the groE genes, and (c) other examples of bacterial gene products involved in phage assembly, to show the extent to which phage utilize the host machinery to successfully carry out their morphogenesis. Function of the groE genes in bacteriophage assembly A head assembly pathway Only the A head assembly pathway seems to be affected in groE­mutants so I will limit my discussion to a description of the normal head assembly process and the effect of growth in a groE- mutant. A genetic map of A emphasizing the morphogenetic genes is 4 shown in Fig. 1. Some of the proteins encoded by this region (e.g., gp~, gp~, gpW, and gpFII) are structural components of the completed virion. Some proteins .are present in modified forms (i .e., some gp~ is cleaved to p!* ' and parts of gp~ and gpC make up p~ and p~,g). The Nu3 gene product participates only transiently in assembly. Head assembly can be divided into five stages (25), during which the phage proteins and DNA molecules proceed from unassembled forms to the stable, expanded, DNA-filled heads capable of attaching to tails (Fig. 2). These stages are only possible explanations of the data and are no means proven to be obligately ordered as shown in Fig. 2. Stage I. The first stage could be the formation of the initiator or preconnector, consisting of about 12 molecules of gpB and various cleaved forms of gp~. The evidence for this structure is biochemical, genetic, and circumstantial. The first multimolecular precursor formed may be a complex of gpgroEL and gp~ (which is the major component of the head-tail connector (26)). These two proteins cosediment at 25S on glycerol gradients (17). The genetic evidence of Georgopoulos et ale (11) (see below) and the cosedimentation of gpgroEL and a gp~ amber fragment (17) suggest that there is a specific interaction between the two proteins. In later stages of infection, gp~ cosediments with cleaved forms of gp~ (called p~l) (17) These gp~-gpgroEL and gpB_p~' complexes are only formed in extracts containing active gpNu3 (the scaffolding protein). This protein may be only transiently required for assembly or it could be loosely attached to those complexes, since it is exclusively found at the top of glycerol gradients. Ferrucci and Murialdo (27) suggested 5 Figure 1. Physical and transcriptional map of the A head gene region. Above is a circular map of A The expanded region is the late transcription unit, which includes the head genes described in the text. 6 r--' \0 1-1 \ -c........t-- o UO \ "'0 \ ~ /:a- ::: ::n \ ~~~ , ::: ~ ::n \ o jN \ )jj \ ::, \ .'. ' IIII n - -'z- 0en c » H i=E CD "'T1 / ...... / Ig~J/ n 0 en I / / / / \ \ I I "'T1 1=1 / / I ..c.. .. ~ {;; Figure 2. Bacteriophage A head assembly pathway. The proteins must enter in the indicated order for proper assembly to occur. The tail joins spontaneously to the mature heads. gpl, gpD, and gpNu3 are written in larger characters to indicate that large numbers of molecules of these proteins participate in head assembly. 8 gpE + gpC ...... pXl, pX2 gpa ...... pa* gpNul gpA toil gpFI gpO gpW gpFn -"" - (f) '\ ~ 0 li l. ~~~ '''I~" N3* . ~ ~ gpC gp groEL gp groES immature proheod p U mature • prohead (unstable) concatemeric DNA head phage that gp! and gpNu3 form a complex, since gpB-donor and gpNu3-donor extracts can only complement gpl-donor extracts (in the in vitro prohead assembly system of Murialdo and Becker (28,29)) if they are concentrated and mixed first. These results are consistent with the possible pathway for preconnector assembly in Fig. 3 (25). The absence of gpgroES from this scheme stems from its recent discovery (30), after most of the above experiments were compl eted. Stage II. The preconnector is thought to serve as a nucleation site for the condensation of 420 gpI molecules to form a shell (31). Several results suggest that gpNu3 forms a core in the prohead. First, there are gpNu3 - containing proheads after infection of wild-type cells (31; C. Georgopoulos, personal communication). Proheads containing gpNu3 are also found after infection of wild-type hosts with £- or B- mutants or infection of groE- mutants with wild-type A (33). Also, Zachary et ale (34) observed, in the electron microscope, proheads with cores, either at the early stages of norl11al infection or with AC- or I. coli groEL­mutants. In the case of infection of wild-type phage and hosts, core-containing proheads were replaced late in infection by empty proheads or filled heads. These results all suggest that gpNu3 forms the cores and that these cores serve a function in prohead assembly. * Stage III. Stage III of head assembly is prohead maturation. Three proteolytic cleavages make up this stage: a. pl (56,000-M r) is formed from about two-thirds of the gpl (62,000-M r) molecules in the head (35), b. The gpNu3 is broken down and leaves the head (36), and c. each gp~ reacts with a gpI molecule to form 10 11 Figure 3. Early steps in A prohead assembly. gpNu3 gP!) gpgroEL ... gpgroEL-gp.!!. - (9PNU3?)---7-.,.....; ........ gpC- gp.!!.(9PNU3?) 9 pC gp_gfoE L the fusion/cleavage products p~ or p~ (35). Stage IV. The order of steps in this stage is unclear. The Nul, ~, and £l gene products facilitate interaction of the prohead with the concatameric DNA, packaging of a genome-length of DNA into , the head, and cutting of the DNA at the cohesive end sites (37). The shell must expand to accomodate the DNA and about 420 molecules of gp.Q. a re added (38). Stage V. The final stage is the addition of gpW and gpFII. These two proteins stabilize the head and make it competent for tail addition (39-41). Position of the groE block. The ~ proteins seem to act very early in A assembly, since some of the first assembly intermediates are abnormal or absent. In the groEL- mutants tested, gp~ still seemed to sediment with the mutated gpgroEL (17). The canplex ;s altered, however, with a lower ratio of gp~ to gpgroEL in the 25S peak. It also appears that the gpB_p~' complex is not formed ;n the absence of at least active gpgroEL (17). Earlier experiments suggested that in vitro prohead assembly systems require that all extracts be prepared from lysates in groE+ bacteria (27). Recent results (J. Kochan, personal communication) suggest that the two groEL- mutants tested (groEL140 and groEL764) affect only preconnector formation. It was shown that preconnectors (the precursors of the head-tail connector, consisting of gp~) made in groE+ lysates can be assembled into infectious phage by in vitro complementation with extracts from infection of groE- hosts, suggesting that at least the particular mutants tested only affect 13 preconnector assembly. It may still be that the groE proteins are required at steps later in assembly but the early block in the known groE- mutants has prec 1 uded ,study of those steps. Some head-related structures are formed in groE- bacteria. Tubular or "monster" assemblies of gpi, were observed upon electron microscopic analysis of lysates from groEL- bacteria (11). Proheads resemb 1 i ng immature proheads, wi th unc 1 eaved gp.l, gpi" and gpl and about 200 molecules of gpNu3 can be isolated after infection of groEL- or groES- bacteria (42-44). These proheads are probably dead­end assembly structures with the low level of proheads that can be matured in vitro (32) stemming from the leakiness of the groE­mutations. The genetic experiments of Georgopoulos et al. (11) also contribute to our understanding of the roles of groE gene products in A head assembly. A mutants (E) were isolated on the basis of their ability to plate on groEL- or groES- bacteria. Some of the mutants isolated on groEL764 had mutations in the.l gene, suggesting that gp~ interacts with gpgroEL. These E mutants grew on some groEL­mutants but not others. The allele-specificity indicates that the mutations could alter protein-protein interactions. A second class of mutants (isolated on both groEL- and groES- hosts) consists mainly of phage with nonsense or temperature-sensitive mutations in the I gene (11). Their ability to grow is probably the result of the lower levels of ~ found under "permissive" conditions (supE, the suppressor tested, suppresses inefficiently (45)). Growth in these conditions could be a result of the IIFloor effect ll (46). If there is 14 an imbalance of components, then the assembly of many structures may be initiated but none completed. Lowering the levels of a particular component (e.g., gpI) can allow the completion of a small number of active structures. A lambdoid phage, HK97, has recently been shown to cross-link its major capsid protein into very high molecular weight molecules (M. Pop a and R. Hendrix, personal communication). This cross-linking was not performed in groEL- bacteria, suggesting that gpgroEL may perform an additional role in HK97 head a·ssembly. It is currently difficult to understand all these observations relating to groE function in terms of a single coherent model for head assembly. One possibility is that gpgroEL serves as a nucleation site for gp~ assembly into preconnectors. The phenotype of groES- mutants for A assembly is the same so gpgroES may be present in such a complex. Biochemical evidence for such a complex, however, is completely lacking at this time. Role of groE in T4 assembly The greater complexity of the T4 head is reflected in the differences between its assembly pathway and that of the A head (see Wood and King (47) for review). Several non-structural proteins (including gp~) act on gp~ (the major capsid protein) and the structural proteins gp~ and gpIPIII before the prohead is formed, in association with the cytoplasmic membrane. gpQ is cleaved to pQ* and gp24 adds to the head in the next stage. The subsequent step, DNA encapsidation, requires the addition and cleavage of several structural proteins and the catalytic action of several more. 15 Finally, more proteins add and some are further modified before the head is mature and able to bind tails. As in A head assembly, the role of the groEL gene product in T4 morphogenesis seems to be very early. Infection of groEL44 bacteria with wild-type T4 leads to the accumulation of gp~ as lumps on the bacterial membrane (48). None of the gp~ is cleaved to p23* and other proteolytic cleavages in the assembly pathway do not occur. This is exactly the phenotype of T411- infection in wild-type bacteria (49). Support for the existence of functional interaction between gpgroEL and gp11 lies in the genetic results of Georgopoulos et al. (48). These authors showed by complementation and recombination that T4 mutants able to grow on groEL44 (T4e:) had mutati ons in gene B-. Other groups have obtained similar results with mutants that are probably groEL-. Takano and Kakefuda (50) also isolated a mutant (which they called mop) which map~ in the groE region and blocks T4 head assembly. They isolated T4 gene 11 mutants that can grow on mop- bacteria. Revel et al. (51) isolated bacterial mutants (hdh) that specifically block T4 productive infection at the level of head assembly. Some map in the groE region and probably affect the groEL gene, since lysogenization with a groEL+ transducing phage makes the bacteria able to propagate T4 (K. Tilly, unpublished results). T4 mutants were selected and mapped (51). Some had mutations in gene 31 but other mutations mapped in gene~. The second class of T4e: mutants may be analogous to the Ae: mutants with mutations in the E 16 gene: mutants with lowered functional levels of the head protein so that a few structures are assembled. These mutants were found about 100-fold more frequently than those with gene 11 mutations and the phage formed small plaques (51), consistent with the idea that they reduce protein levels, rather than altering a specific interaction. Similar results were obtained by Coppo et al. (52) and Takahashi et ale (53) with their independently isolated host mutants. So far, no groES- mutants have been found that affect T4 growth. This could be a consequence of the smaller target size of the gene but it could be that T4 only requires gpgroEL for head assembly (see Chapter V). T5 Tail Assembly Little is known about the process of T5 tail assembly. The seven protein components of the complete tail have been identified by electrophoresis of proteins from purified tails on polyacrylamide gels (54). Pulse-chase experiments have shown that one of the proteins is cleaved during assembly (55). Infection of some groEL­or groES- mutants with T5 results in lysates that will donate normal numbers of functional heads, although there are more tubular forms and polyheads than in a normal infection (55). The protein cleavages found in T5 head assembly do take place. These lysates, however, do not contain significant levels of tails active in in vitro complementation assays. The cleavage normally found in tail assembly is not performed. T5£ mutants have been isolated and one mutation has been mapped to either the major tail protein gene or the gene encoding the protein that is cleaved (55). 17 Summa ry of groE- effects on phage assembly Several common elements are found in the phenotypes of T4, T5, and A infection of groE- hosts. First, the block seems to be at a very early step in assembly. (Not enough is known about T5 tail assembly to be sure that the block is early in that case.) No maturab1e structures are formed (with the possible exception of A, which may be simply the result of a leaky ~ mutation (32)). Second, some proteolytic cleavages fail to occur in all cases. Third, it is possible to isolate phage suppressor mutants, called E, with alterations in genes whose products are known to participate in the very early stages of assembly. These lines of evidence point toward an early role for the groE proteins in the assembly of many kinds of phage. In fact, there may be an analogue to gpgroEL in the Gram-positive bacterium Bacillus subti1is, since an antigenical1y­related protein is found associated with q, 29 proteins early in infection (J. Carrascosa, personal communication). Effect of groE- mutations on bacterial growth Many groEL- or groES- bacterial mutants are temperature­sensitive for growth, indicating that the groE gene products are essential for the bacteria (11). Bacteria shifted to 42°C exhibit several phenotypes in addition to their inability to form colonies. The mutants form long filaments without septa (55), suggesting the existence of a direct or indirect effect on septum formation. The filamentation is reversible for at least a few hours, since a down­shift to 30°C of a 42°C bacterial culture results in septum 18 formation, cell division, and an increase in viable counts (56). Bacterial filamentation is accompanied by altered patterns of protein synthesis (56). Some proteins are made at higher levels while others are made at lower levels after shifting to 42°C. One protein whose synthesis is shut off is membrane-associated (55) but the functions of this and the other regulated proteins have not been identified. Temperature-resistant revertants of all groE mutants tested simultaneously acquired the ability to plate ~, normal cell division at high temperature, and normal protein synthesis patterns at 42°C, indicating that all of these phenotypes result from single mutations. Sternberg (57) found that ribosomal RNA and tRNA levels were reduced significantly in one groE- mutant (isolated by a different selection) at 44.5°C. He observed that temperature-resistant revertants simultaneously regained the ability to plate ~ and to synthesize wild-type rRNA and tRNA levels. He also found, however, that transduction of the groE- mutation into a new strain background (yielding temperature-sensitive bacteria blocking ~ head assembly) led to a loss of the effect on RNA levels. This suggests that at least one unlinked mutation was required, in addition to the groE­mutation, for this phenotype. groE gene regulation Analysis of conditions that induce or repress groE protein and mRNA synthesis has provided circumstantial evidence about the role of the gene products in bacterial growth. Many stimuli increase the synthesis rates of the groE gene products relative to the rest of the 19 I. coli proteins. The best-studied stimulus is heat-shock. Heat­shock in Drosophila melanogaster causes a dramatic reduction in the transcription of most genes while a small group of genes is induced (see Ashburner and Bonner (58) for review). There also appears to be translational control so that heat-shock mRNA molecules are preferentially translated. Many other eukaryotic organisms have been shown to have similar responses (i.e., induction of a few proteins and shut-off of the majority) but the mechanisms and kinetics of the control vary (see for example (59)). One function of the heat-shock response in Chinese hamster fibroblasts is to provide protection from thermal killing (60) but others probably remain undiscovered. The heat-shock response in I. coli is quite different from that in eukaryotes. A temperature shift-up results in a transient large increase in the rate of synthesis of a group of proteins, followed by a decrease in the rates to close-to-normal levels (61,62). Most other proteins continue to be synthesized at their normal rates. Recent experiments (63; 64) have shown induction of the heat-shock genes to be controlled by a single locus. A bacterial strain with an amber mutation in the htpR (63) or hin (64) gene (which maps at 75 minutes) was isolated as a temperature­sensitive mutant in a strain carrying a temperature-sensitive amber suppressor (65). This mutant became temperature-resistant when made lysogenic for a Q> 80supF transducing phage, indicating that the mutation was an amber mutation in an essential gene (65). Analysis of the overall protein synthesis patterns in this mutant after a shift to 42°C (at which temperature the temperature-sensitive 20 suppressor is inactive) revealed that it fails to induce the synthesis of the previously characterized heat-shock polypeptides (63,63). Careful analysis showed this group of proteins to consist of at least nine polypeptides (63). Four of these proteins have been identified as gpgroEL, gpgroES, gpdnaK (a protein required for A DNA replication and l. coli DNA and RNA synthesis (8)), and form II of lysyl tRNA synthetase (66). The other proteins have only been identified as spots on two-dimensional polyacrylamide gels (63,64). Since htpR- cultures lyse after growth at 42°C (63), it seems likely that at least one of the heat-shock polypeptides is required for survival at that temperature. Yamamori and Yura (64) have studied the level at which the groE genes are regulated by the ~ gene. They found seven-fold heat induction of RNA that will hybridize in liquid to a groE+ transducing phage. In the htpR mutant, there was no more than a two-fold induction of such RNA, suggesting that htpR control is exerted at the level of transcription. Yamamori and Yura (64) also addressed the question of what role the heat-shock polypeptides play in bacterial growth. They showed that the growth rate of the htpR- mutant rose coordinately with the activity of the amber suppressor present in the strain. They also found that incubation of wild-type cells at 42°C for 30 minutes resulted in a 10 to 100-fold increase in resistance to thermal killing at 55°C, while htpR- mutants were unprotected by a 42°C treatment. Growth at 42°C and protection from thermal killing require at least one of the heat-shock proteins. It would be logical that only 21 genes with a selective advantage would be maintained under heat-shock control so it is likely that both of the ~ gene products are important in these two processes. Induction of the heat-shock proteins through ~ seems to be caused by other stresses on the cell. Recent experiments have examined the effect of A infection on the synthesis of I. coli proteins (67,68). It was found that gpgroEL and gpdnaK (among others) are transiently induced by A infection. It is not clear from these data whether or not gpgroES synthesis is induced but the results of Georgopoulos (see below) suggests that the induction requires an active htpR gene product. Drahos and Hendrix (67) and Kochan and Murialdo (68) analyzed the effects of various deletion­substitution derivatives of A in order to pinpoint the A gene or genes responsible for the induction and obtained somewhat conflicting results. The effect clearly comes from the PL operon of A (whose genes are non-essential for lytic growth) and more than one gene may have an effect. There is clear coordinate induction of the groE genes after infection of A~I857cr027 phage into ~+ bacteria at 42° (the Tro phenotype; C. Georgopoulos, personal communication). The phage yield in these conditions is reduced more than lOO-fold from that at 37°C and overall bacterial DNA, RNA, and protein synthesis is drastically reduced. The effects seem to result from overproduction of products from the PL operon, most likely gpEalO and gp~ (69). Analysis of the protein synthesis patterns after shifting a A£I857cr027 lysogen to 42°C for various periods of time shows that synthesis of the majority 22 of bacterial proteins is shut off but synthesis of a few persists at induced levels for many hours after the shift (C. Georgopoulos, personal communication). All early phage proteins (notably gp~10) are also overproduced. The I. coli proteins induced are those controlled by gphtpR and induction is not found in htpR- bacteria. A may bias most host protein synthesis early in infection toward proteins that it requires for growth. Later on, this effect could be repressed by gpcro (presumably turning off PL transcription (70)). This shut-off would be defective in cro- phage, leading to the continued high rate of translation of PL-transcribed genes and, therefore, of the heat-shock proteins. Induction by gphtpR is also found when the level of RNA polymerase subunit a is lowered (71). Low a levels were achieved by growing an I. coli strain with a a am mutation (rpoD40; 72) and a temperature-sensitive ~ allele at temperatures at which the suppressor has varied activity levels. It was found that at least the three largest heat-shock proteins (including gpgroEL) were made at high levels even 90 min. after a shift to 42°C, at which time the heat-shock response has subsided (61). This phenotype was not found ina a am htpR am double mutant ina supFts background, confi rmi ng that gphtpR is again required for the induction of heat-shock polypeptide synthesis. A final case of groE induction is in the stringent response, which results from increased levels of ppGpp after amino acid starvation. This leads to a rapid and dramatic decrease in overall E. coli protein synthesis (73). The expression of some genes, 23 however; is stimulated at the transcriptional level (74). Analysis of the proteins made after partial amino acid starvation has clearly shown that gpgroEL is one of the last proteins whose synthesis is shut off during starvation (73). Chao (74) found that a 15,OOO-Mr polypeptide also persists. This protein may be gpgroES. C. Woolford (personal communication) has found continued gpgroES expression after amino acid starvation. These results suggest that the ~ proteins may playa role either in the stringent response, in the survival of starving bacteria, or in both. In summary, the groE genes seem to be induced when the cell is under stress. In many cases, this induction seems to be a consequence of gphtpR action. Induction by amino acid starvation may be the result of a different mechanism since it is not clear whether or not the other gphtpR-controlled genes are induced. A probably turns on the htpR pathway, but not necessarily because all the products are important to A. Alternatively, the host may perceive infection as a stress, causing it to induce the htpR-controlled genes. Other host genes involved in phage assembly I. coli gene products other than those of the groE genes clearly play roles in bacteriophage head assembly, tail assembly, and DNA packaging (75). There is both genetic and physical evidence for such interactions. Mutations altering the lipid composition of the bacterial inner membrane (fatA; 76) and the timing of T4 late gene expression (probably in the rho gene; 77-81 and possibly 82) affect T4 head assembly. E. coli mat mutants were isolated because both - ---- 24 head and tail assembly are defective upon infection. but a stage before late gene expression of the closely-related phage 434 is blocked in such strains (83; C. Waghorne et al., personal communication). Heat treatment of the host before infection with a T4regA- mutant lowers the level of abnormal head structures (84), suggesting that more host factors are involved. Unidentified host proteins are found associated with many phage head structures (17, 85,86; C. Georgopoulos, personal communication) but their significance is unknown. These apparent interactions could be fortuitous or functional. Bacterial mutants either bypassing some phage functions required for T4 tail fiber assembly or preventing that process have been isolated (52,87,88). There also seems to be interaction between T4 baseplate proteins and host components (89,90). Bacteria also seem to play roles in phage DNA packaging. 21, a lambdoid phage, cannot grow on himA-strains (which were isolated as mutants in which l integration is defective; 10). Phage mutants able to grow on himA bacteria have mutations in the Nul gene (which encodes a terminase subunit; (91,92; S. Frackman and M. Feiss, personal communication). M. Gold et ale (personal communication) have shown that purified terminase preparations require at least two host proteins (neither of which is gphimA) in order to cleave cos sites in vitro. Host nucleic acid modification enzymes seem to be required for packaging. T4 infection of end-polA- bacteria is abnormal (93). These enzymes may be involved in resolving the branched DNA molecules 25 formed late in T4 DNA replication into packageable structures. Host ligase and T4 ligase also seem to work synergistically in preparing the DNA for packaging (94). Infection of T7 into E. coli tsnS-mutants (which have mutations in the 8 or 8' subunit of RNA polymerase) leads to the accumulation of shorter than unit length DNA molecules, probably because a defect in packaging makes the DNA sensitive to exonucleases (95). Finally, there are host mutants in which filamentous phage extrusion (which combines packaging and assembly) is defective (96). In summary, it appears that the host participates in phage assemb lyon many 1 eve 1 s. The host memb rane se rves as an anchor on which some phages build structures. Host proteins are required for head assembly, tail assembly, and DNA packaging and are able to substitute for some phage-encoded products. The complex network of interactions is probably only a small portion of the total. In the following chapters, I will describe an approach to understanding these interactions, by the genetic and physical analysis of two E. col i genes requi red for A assembly. 26 Literature Cited 1. Lin, L., Bitner, R. & Edlin, G. (1977) J. Viral. 21:554-559. 2. Kornberg, A. (1980) DNA Replication (W.H. Freeman and Company, San Francisco), pp. 473-569. 3. Luria, S.E., Darnell, J.E. Baltimore, D. & Campbell, A. (1978) General Virology (John Wiley & Sons, New York), pp. 157-250. 4. Falco, S.C., Vander Laan, K. & Rothman-Denes, L.B. (1977) Proc. Natl. Acad. Sci. USA 74:520-523. 5. Mount, D.W.A., Harris, A.W., Fuerst, C.R. & Siminovitch, L. (1968) Virology ~:134-149. 6. Scandella, D. & Arber, W. (1976) Virology 69 206-215. 7. Georgopoulos, C.P. (1971) in The Bacteriophage Lambda ed. Hershey, A.D. (Cold Spring Harbor Laboratory, Cold Spring Harbor), pp. 639-645. 8. Saito, H. & Uchida, H. (1977) J. Mol. Biol. 113:1-25. 9. Miller, H.I. (1981) Cell 25:269-276. 10. Georgopoulos, C.P., Hendrix, R.W., Casjens, S.R. & Kaiser, A.D. (1973) J. Mol. Biol. ~:45-60. 11. Georgopoulos, C.P. (1971) Proc. Natl. Acad. Sci. USA 68:2977- 2981. 12. Bachmann, B.J. & Low, K.B. (1980) r~icrobiol. Rev. 44:1-56. 13. Guest, J.R. & Nice, H.M. (1978) J. Gen. Microbial. 109:329-333. 14. Tabor, H., Hafner, E.W. & Tabor, C.W. (1980) J. Bacterial. 144:952-956. 15. Georgopoulos, C.P. & Hohn B. (1978) Proc. Natl. Acad. Sci. USA 75:131-135. 16. Hendrix, R.W. & Tsui, L. (1978) Proc. Natl. Acad. Sci. USA 75:136-139. 1 7 • Mu ria 1 do, H • (1979) V; ro logy 1§.: 341-367 • 18. Hendrix, R.W. (1979) J. Mol. Biol. 129:375-392. 19. Hohn, T., Hohn, B., Engel, A., Wurtz, M & Smith, P.R. (1979) J. Mol. Biol. 129:359-373. 27 20. Fuchs, E., Zillig, W., Hofschneider, P.H. & Preuss, A. (1964) J. Mol. Bio1. 10:546-550. 21. Lubin, M. (1969) J. Mol. Bio1.12..:219-233. 22. Ishihama, A., Ikeuchi, T., Matsumoto, A.& Yamamoto, S. (1976) J. Biochem 79:927-936. 23. Paetkau, V. & Coy, G. (1972) Can. J. Biochem. 50:142-150. 24. Ishihama, A., Ikeuchi, T. & Yura, T. (1976) J. 8iochem ~:917- 925. 25. Georgopoulos, C., Tilly, K. & Casjens, S. (1982) in The Bacteriophage LambdaII, eds. Hendrix, R., Weisber~R., Stahl, F & Roberts, J. (Cold Spring Harbor Laboratory, Cold Spring Harbor), in press. 26. Tsui, L. & Hendrix, R.W. (1980) J. Mol. Biol. 142:419-438. 27. Ferrucci, F. & Muria1do, H. (1981) in Pro ress in Clinical and Bio10 ica1 Research, ed. Dubow, M. A an R. Liss, Inc., New York Vo1.6, pp. 193-212. 28. Murialdo, H. & Becker, A. (1977) Proc. Nat1. Acad. Sci. USA 74:906-910. 29. Muria1do, H. & Becker, A. (1978) J. Mol. Bio1. 125:57-74. 30. Tilly, K., Muria1do, H. & Georgopoulos, C. (1981) Proc. Natl. Acad. Sci. USA 78:1629-1633. 31. Muria1do, H. & Becker, A. (1978) Microbiol. Rev. ~:529-576. 32. Hohn, T., Flick, H. & Hohn, B. (1975) J. Mol. Bio1. 98:107-120. 33. Georgopoulos, C.P., Bisig, R., Magazin, M., Eisen, H. & Court, D. (1979) J. Viro1. 29:782-788. 34. Zachary, A., Simon, L.D. & Litwin, S. (1976) Virology 72:429- 442. 35. Hendrix, R.W. & Casjens, S.R. (1975) J. Mol. Biol. 91:187-199. 36. Ray, P. & Murialdo, H. (1975) Virology i1,:247-263. 37. Feiss, M. & Becker, A. (1982) in The Bacteriophage Lambda II, eds. Hendrix, R., Weisberg, R., Stahl, F. & Roberts, J. (Cold Spring Harbor Laboratory, Cold Spring Harbor), in press. 38. Imber, R., Tsugita, A., Wurtz, M. & Hohn, T. (1980) J. Mol. Biol. 139:277-295. 28 39. Casjens, S.R., Hohn, T. & Kaiser, A.D. (1972) J. Mol. Biol. 64:551-563. 40. 41. 42. Casj ens, Katsura, Ku·nzl er, s. I. P. ( 1974) J. Mo 1. B i 01. .2Q.: 1-2 3 • & Tsugita, A. (1977) Virology 76:129-145. & Hohn, T. (1978) J. Mol. Bi 01 • 122:191-215. 43. Earnshaw, W.C., Hendrix, R.W. & King, J. (1979) J. Mol. Biol. 134:575-594. 44. Tilly, K., McKittrick, N., Georgopoulos, C. & Murialdo, H. (1981) in Pro ress in Clinical and Biolo ical Research, ed. Dubow, M. Alan R. Liss, Inc., New York, Vo • 64, pp. 35-46. 45. Garen, A. (1968) Science, 160:149-159. 46. Floor, E. (1970) J. Mol. Biol. 47:293. 47. Wood, W.B. & King, J. (1979) -in Com rehensive Virology eds. Fraenkel-Conrat, H. & Wagner, R.R. Plenum Publishing Corp., New York), Vol. 13, pp. 581-633. 48. Georgopoulos, C.P., Hendrix, R.W., Kaiser, A.D. & Wood, W.B. (1972) Nat. New Biol. 239:38-41. 49. laemmli, U.K., Beguin, F. & Gujer-Kellenberger, B. ( 1970) J. Mol. Biol. 47:69-85. 50. Takano, T. & Kakefuda, T. (1972) Nat. New Biol. 239:34-37. 51. Revel, H.R., Stitt, B.l., Lielausis, I. & Wood, W.B. (1980 ) Virol.33:366-376. 52. Coppo, A., Manzi, A., Pulitzer, J.F. & Takahashi, H. (1973) Mol. Biol. 76:61-87. 53. Takahashi, H., Coppo, A., Manzi, A., Martire, G. & Pulitzer, J.F. (1975) J. Mol. Biol. ~:563-578. 54. Zweig, M. & Cummings, D.J. (1973) Virology ~:443-453. 55. Zweig, M. & Cummings, D.J. (1973) J. Mol. Biol. 80:505-518. J. J. 56. Georgopoulos, C.P. & Eisen, H. (1974) J. Supramol Struct. 2:349- 359. 57 • S t ern b erg, N • (1973) J. Mo 1. B i 01. 76: 1-2 3 • 58. Ashburner, M. & Bonner, J.J. (1979) Cell 17:241-254. 29 59. Lindquist, S. (1981) Nature 293:311-314. 60. Li, G.C. & Werb, Z. (1982) Proc. Natl. Acad. Sci. USA 79:3218- 3222. 61. Yamamori, T., Ito, K., Nakamura, Y. & Yura, T. (1978) J. Bacteriol. 134:1133-1140. 62. Lemaux, P.G., Herendeen, S.l., Bloch, P.l & Neidhardt, F.C. (1978) Cell ~:427-434. 63. Neidhardt, F.C. & VanBogelen, R.A. (1981) Biochem. Biophys. Res. Commun. 100:894-900. 64. Yamamori, T. & Yura, T. (1982) Proc. Natl. Acad. Sci. USA 79:860-864. 65. Cooper, S. & Ruettinger, T. (1975) Mol. Gen. Genet. 139:167-176. 66. Hirshfeld, I.N., Bloch, P.l., VanBogelen, R.A. & Neidhardt, F.C. (1981) J. Bacteriol. 146:345-351. 67. Drahos, D.J. & Hendrix, R.W. (1982) J. Bacteriol. 149:1050-1063. 68. Kochan, J. & Murialdo, H. (1982) J. Bacteriol. 149:1166-1170. 69. Georgiou, M., Georgopoulos, C.P. & Eisen, H. (1979) Virology 94:38-54. 70. Eisen, H., Brachet, P., Pereira da Silva, l. & Jacob, F. (1979) Proc. Natl. Acad. Sci. USA 66:855-862. 71. 72. 73. 74. 75. 76. 77. Osawa, T. & Yura, T. (1981) Mol. Gen. Genet. 184:166-173. Osawa, T. & Yura, T. (1980) Mol. Gen. Genet. 180:293-300. Reeh, S., Pedersen, S. & Friesen, J.D. (1976) Mol. Gen. Genet. 149:279-289. Chao, l. (1977) Biochem. Biophys. Res. Commun. 77:42-49. Georgopoulos, C. & Tilly, K. (1981) in Progress in Clinical and Biological Research, ed. Dubow, M. (Alan R. liss, Inc., New York), Vol. 64, pp. 21-34. Simon, l.D., Mclaughlin, J.J.M., Snover, D., Ou, J., Grisham, C. & loeb, M (1975) Nature 256:7379-7383. Simon, l.D., Snover, D. & Doermann, A.H. (1974) Nature 252:451- 455. 30 78. Simon, L.D., Gottesman, M., Tomczak, K. & Gottesman, S. (1979) Proc. Nat1. Acad. Sci. USA 76:1623-1627. 79. Takahashi, H. (1978) Virology ~:256-265. 80. Caruso, M., Coppo, A., Manzi, A. & Pulitzer, J.E. (1979) J. Mol. Biol. 135:959-977. 81. Stitt, B.L., Revel, H.R., Lielausis, I. & Wood, W.B. (1980) J. Virol. 35:775-789. 82. Jensen, J.L. & Susman, M. (1980) Genetics ~:301-325. 83. Fuerst, C.R., Bingham, H. & Bouchard, J.P. (1978) Virology 87:416-436. 84. Wever, G.H., Thompson, B.J., Laiker, R.M., Ruby, E. & Wiberg, J.S. (1981) in Pro ress in Clinical and Biolo ical Research, ed. Dubow, M. (Alan R. Liss, Inc., New York, Vol. 64:167- 192. 85. Onorato, L. & Showe, M.K. (1975) J. Mol. Biol. ~:395-412. 86. Earn shaw, W. & Kin g , J. ( 1978) J. Mo 1. B i 01. 126 : 721-747 • 87. Revel, H.R., Herrman, R. & Bishop, R.J. (1976) Virology 11:255- 265. 88. Pulitzer, J.F. & Yanagida, M. (1971) Virology ~:539-554. 89. Simon, L.D. (1969) Virology l.§.:285-296. 90. Georgopoulos, C., Georgiou, M., Selzer, G. & Eisen, H. (1977) Experientia ~:1157-1158. 91. Sumner-Smith, M. & Becker, A. (1981) Virology 111:629-641. 92. Sumner-Smith, M., Becker, A. & Gold, M. (1981) Virology 111:642- 646. 93. Chao, J., Chao, L. & Speyer, J.F. (1974) J. Mol. Biol. ~:41-50. 94. Black, L.W., Zachary, A.L. & Manne, A. (1981) in Progress in Clinical and Biological Research ed. Dubow, M. (Alan R. Liss, Inc., New York), Vol. 64, pp.111-126. 95. DeWyngaert, M.A. & Hinkle, D.C. (1980) J. Virol. ~:780-788. 96. Model, P., Russel, M. & Boeke, J.D. (1981) in Progress in Clinical and Biolo ical Research, ed. Dubow, M. (Alan R. L1SS, Inc., New Yor ,Vol. 6 , pp. 389-400. 31 ERRATUM The following page did not receive a number. CHAPTER II IDENTIFICATION OF A SECOND ESCHERICHIA COLI GROE GENE WHOSE PRODUCT IS NECESSARY FOR BACTERIOPHAGE MORPHOGENESIS* Introduction The Escherichia coli groE locus codes for a major cellular protein that has been shown to be required for a wide variety of morphogenetic processes. GroE- mutants, although selected as bacteria that prevent productive infection by bacteriophage A are mutated in an essential bacterial function because many are temperature sensitive for growth, forming long filaments without septa at 42°C (1). Pleiotropic effects of these mutations include blocks on T4 head morphogenesis and T5 tail assembly (2,3). Wild­type phage A infecting groE- bacteria are blocked in head morphogenesis (4,5) because only abnormal head structures are formed and none of the known proteolytic cleavages required for normal head assembly takes place (6,7). This;s consistent with the finding that the gene product of groE participates in an early step in A prohead *This chapter is an exact copy of Tilly, K., Murialdo, H. & Georgopoulos, C. (1981) Proc. Natl Acad. Sci. USA~, 1629-1633. assembly, in which the groE product forms a complex with the gene product of ! (7). The groE- mutations have been mapped at approximately 93 mi n on the i.. col i map and are analogous to mutations called mop (8), and hdh (10). The bacterial groE locus has been cloned by in vitro recombination into several A vectors (7,11,12), which has allowed the identification of the gene product of ~ as a polypeptide of 65,000 Mr - The purified polypeptide is found as a decatetramer sedimenting at 25S, its 14 subunits arranged with 7-fold symmetry (13,14). A weak ATPase activity is associated with the complex, but the normal function of this protein in uninfected bacteria is not known. In this paper, we present a genetic and biochemical analysis of phage mutants from which we conclude that mutations in a second bacterial gene also cause the GroE phenotype. We have identified the product of this gene as a polypeptide of 15,000 Mr. We propose to rename the two groE genes groEL (which codes for the sYnthesis of the 65,000 Mr polypeptide) and groES (which codes for the synthesis of the 15,000 Mr polypeptide). According to this new designation, mutants originally named groEA44 and groEB764 (4) will be referred to as groEL44 and groEL764, respectively. Methods and Materials ! Bacterial and phage strains Most bacterial and phage strains used in these studies have been described (1,2,7,11). The term ~+ denotes nonsuppressing bacteria (formerly known as supo, sup- or ~-). Phage H18a is an iA c160 derivative of phage H18 (11). Phage H18 is an 21groE+att+ 32 transducing phage that was constructed in vitro by the insertion of the Hi nd I I I i.. co 1 i DNA fragment carryi ng the groE+ genes into phage vector 540 (15). Phage 375 was constructed in vitro by the insertion of the EcoR I i.. coli DNA fragment carryi ng the g roE+ genes into phage vector A gtB (7,16). Phage 378 is identical to phage 375 except that the orientation of the EcoRI I. coli DNA fragment has been reversed (Fig. 1). Media and bacterial and phage platings were as described (4,7,11). Isolation of phage-deletion derivatives Deletion mutants of groE+ transducing phages were isolated after EDTA treatment (17). Approximately 109 phage were treated with 2 mM EDTA for 1 hr at 42°C. Inactivation was stopped by the addition of MgC1 2 to 0.17 M. The surviving phage were plated on B178 groE+ bacteria on Tryptone plates supplemented with 0.6 mM EOTA. Under these conditions, only phage which have undergone substantial deletion of DNA will form large plaques. This step is necessary because EDTA selection in liquid does not altogether eliminate phage with a full complement of DNA. Surviving phage were subsequently tested for growth on a groES- mutant (groES30) and on a groEL- mutant (groEL140). Isolation of H18agroELam1 The level of mutants in phage H18agroE+ was raised by growth on mutO bacteria (18) in the presence of 5 ug of thymidine per ml; this phage stock was adsorbed onto groEL140supE bacteria and plated on a 1:2 mixture of groEL140~ and grOEL140~+ bacteria, 33 Figure 1. Physical and genetic maps of phage A vectors used in this work. The wild-type genome is divided into 100 units. Heavy lines indicate the position of the groE+ bacterial DNA. Gaps indicate deleted DNA. Arrows mark positions at which the indicated restriction enzymes cleave. Maps: a, wild-type bacteriophage A ; b, W3; c, H18a; d, 375; and e, 378. 34 o. A J attA immA Q 'Q I I I I I I o 20 40 60 80 100 b. W3 EcoRI t t i nin5- -- c HI80 EcoRI t t t t · Hindm ~ ~ ~ d.375 EcoRI t t nin5 BamHI ~ ~ ~ ~ e.378 EcoRI t t nin 5 BamHI ~ ~ ~ + + respectively. After overnight growth at 37°C, small turbid plaques were further tested for growth on B178groE+ and groEL140sup+ bacteria. Phage that made plaques on the former but not on the latter bacteria were tested for growth on roEL140supE. One mutant, H18agroELam1, that fulfilled these criteria for possessing an amber mutation in the groEL gene was found among approximately 30,000 phage progeny grown on mutD bacteria and tested. Labeling experiments The bacterial strains, media, and procedures for labeling with either 35s04-2 or [35S] methionine have been desciibed (11). The procedure for labeling phage-infected, UV-irradiated bacteria with a mixture of 3H-labeled amino acids (3H-amino acids) (20 uCi/ml; 1 Ci = 3.7 x 1010 becquerels) was identical to the conditions for labeling with [35S] methionine, except that no unlabeled amino acids were added to the M9 medium. 35s04-2, [35S] methionine, and the mixture of the 3H-amino acids were purchased from New England Nuclear (NET- 250). Phage DNA preparation, restriction enzyme digestion, ligation, and agarose gel electrophoresis These techniques were described (7,19). Likewise, NaDodS04/polyacrylamide gel electrophoresis was as described (11). Two-dimensional gel electrophoresis The technique used was essentially that described by O-Farrell (20). The first dimension consisted of isoelectric focusing to 36 equilibrium (6400 V/hr) in 1.6%. (wt/vol) and 0.4% Ampholine mixtures (pH 5-7 and 3-10, respectively) in a 4% (wt/vol) polyacrylamide gel. The second dimension was run in NaDodS04/12.5% polyacrylamide gels. Prepa rat i on of A proheads The procedure for isolating A proheads labeled with 35S04-2 was as described by Hendrix and Casjens (6). Results W3a mutant W3 is a AgroE+ transducing phage that carries, in its insert of I. coli DNA, all genes necessary for growth on groE- hosts. We have reported the isolation of a mutant of W3, called W3a, which was no longer able to propagate on groEL743 bacteria (11). When this mutant was tested on other groE- strains, it was found to divide the strains into two classes by its plating behavior (Table 1). Class I groE- bacterial mutants did not propagate W3 , whereas class II groE­bacterial mutants did. Because the a mutation affects the electrophoretic mobility of the 65,000 Mr protein (11), it was presumed to reside in the groE structural gene. We suggested two complementation models to explain the plating properties of W3 phage. The first model assumed that the 65,000 Mr protein is the only groE product and that the ability of W3 to plate on class II groE- bacterial mutants is due to intragenic complementation between the mutant subunits synthesized by W3a and the mutated host groE subunits. The fact that the purified form of the groE gene product 37 Table 1. Plating properties of groE+ transducing phage and their derivatives Phage strains* 116t 113t 1113t Host E. coli - -- (grou (grou (grou s tra i n A W3 W3a H18a pl ) p2) p3) H18agroELaml 378 B178groE+ + + + + + + + + groEL140sup+ + + + groEL140supE, F + + + + groES30sup+ + + + + + + groES30supF + + + + + + * +, an efficiency of plating of 0.5-1.0; -, an efficiency of plating of <10- 3 at 37°C. The efficiency of plating of B178 groE+ at 37°C is taken as 1.0. + + + + + t Phages 116, 113,and 1113 are deletion derivatives of H18a, whereas phage 11101 is a deletion derivative of 378. 1l101t (grou p4) + + + w co is a decatetramer (13,14) is consistent with such a model. The second model, intergenic complementation, was that there are at least two groE genes, and mutations in either one can cause the same bacterial phenotype. According to this model, W3a is a mutant in the groEL gene coding for the 65,000 Mr polypeptide and is able to grow on class II groE- bacterial mutants because class II mutants have a normal groEL gene but are mutated in a second groE gene, designated groES. Because both the temperature-sensitive phenotype for bacterial growth at 42°C and the abnormal A prohead structures formed at 37°C are similar for groE- mutants in either class (Fig. 2A; unpublished data), these characteristics could not be used to distinguish between these models. Isolation of a groEL amber mutant To distinguish between the possibilities of intragenic and intergenic complementation, we searched for suppressible amber mutants in the groEL gene, which would be less likely to form functional hybrid complexes. One such mutant, called groELam1, was isolated after mutagenesis of H18agroE+ transducing phage by growth on mutD bacteria. This mutant was selected on the basis of its ability to plate well on groE+~+ and groES30sup+ bacteria, but not on groEL140~+ bacteria (Table 1). When tested further on our other groE- strains, the growth pattern of phage H18agroELam1 on ~+ hosts was identical to that of W3a (Table 1). That is, it plated well on groE+~+ and groES-~+ bacteria, but did not grow on groEL-~+ bacteria. That the groELam1 mutation is indeed suppressible was shown by the ab; 1 i ty of the phage to propagate on groEL-~ or -supF 39 Figure 2. Autoradiograms of NaDodS04/polyacrylamide slab gel electrophoresis. (A) A proheads produced after the induction of Ac1857Sam7 lysogens of B178 groE+ (lane 1) groES7 (lane 2) and groEL44 (lane 3) bacteria. The proheads were labeled with 35S04-2 and separated in glycerol gradients as described (6). Wild-type A proheads are composed of Xl and X2 and gene products of!,!* ,land Q, whereas proheads produced on groE hosts are composed of gene products of £, 1, ~3, and!. (B) UV-irradiated 159uvrA-~+ bacteria infected with various phage and labeled with [35S]methionine (20 uCi/ml) for 10 min at 37°C. Lanes: 1,\£160; 2, H18a; and 3, H18agroELam1. The top arrow points to the 65,000 Mr gene product, which migrates as a doublet, and the bottom arrow points to the putative groELam1-coded amber fragment. (C) UV-irradiated YmelsupF bacteria infected with the same phages as above. The acrylamide was 12.5%. 40 u CD CDI UI WI ~z l a. a. a. a. ~'J( T 1 ~~- ).~ )~ CDI I CDI Wi N - a. a.a.xX c::7' 0' 0' ECCLES HEALTH SC r. t . I 01 0.. 0"\ r<> N r<> N bacteria (Table 1). This was verified by examining the proteins encoded by phage carrying the groELam1 mutation. This was done by infecting UV-irradiated bacteria, labeling the proteins with radioactive precursors, and displaying them on NaDodS04/polyacrylamide gels. It was found that the 65,000 Mr protein band was absent in ~+ infections (Fig. 2B, lane 3) but was present when ~ bacteria were used as hosts (Fig. 2C, lane 3). A potential amber fragment of 35,000 Mr can be seen upon infection of sup+ bacteria by groELam1 phage (Fig. 2B, lane 3). Because phage groELam1 grows on groES-~+ mutants, although it synthesizes only a fragment of gene product groEL, it seems unlikely that its ability to grow is due to intragenic complementation. Isolation of deletion derivatives of H18a To facilitate work with the various phage derivatives, we used the groE+ transducing phage, H18a, which is attA+ (Fig. 1). Phage H18a was treated first with EDTA in liquid and then plated on EDTA­containing plates. Under these conditions, only phage that have suffered a significant loss of DNA will form plaques (17). Upon further testing, the H18a deletion derivatives fell into three groups. In all ,cases, the position and extent of the deletions found in each group have been analyzed genetically (see below) by DNA restriction mapping (Fig. 3) and by DNA heteroduplexing experiments (unpublished results). Group I phage, exemplified by H18ad6, grow on all our groE- bacterial mutant strains. The majority of group 1 deletions originate at the attachment site and delete phage DNA sequences to the right (Fig. 3, lane 3). Group 2 phage, exemplified 42 Figure 3. EcoRI digests of various phage DNAs. Lane 1, A~I60. The fragments are named A,B,C,D,E, and F in order of decreasing size. Lane 2, H18a. This DNA has a new fragment generated by the bacterial DNA insertion, which replaces A fragments o and E. Lane 3, H18a 6. This DNA is deleted from the attA site to the right, fusing the new H18a fragment with the A immunity­containing B fragment. Lanes 4 and 5, H18a63 and H18a613, respectively. These DNAs are deleted from the attA site to the left ending in th~ bacterial DNA substitution, thus reducing the size of the new H18a fragment. 43 F-I 234 5 by H18a~3, do not propagate on any groE- bacterial strains tested. Group 2 deletions originate at the attachment site but proceed to the left, eliminating both phage DNA and bacterial DNA sequences, including the groE genes (Fig. 3, lane 4). Group 3 phage, exemplified by H18a~13, have deletions that also originate at the attachment site and proceed to the left but end in the middle of the groE DNA sequences (Fig. 3, lane 5). We came to this conclusion because H18a~13 phage grow well on all groES- bacterial mutants (on which W3a phage propagate) but do not grow on the groEL- bacterial mutants. The correlation of growth of group 3 deletion phage and W3 is perfect (Table 1): among 21 groE- bacterial mutants tested, all those able to plate group 3 phage also plated W3n. Conversely, all the bacterial mutants unable to plate group 3 phage were also unable to plate W3n. Group 3 phage deletions account for only a few percent of all deletions found. This is not surprising, because they must terminate within the groE DNA region in order to eliminate one but not both of the groE activities. By restriction mapping the extent of the DNA deleted in representative group 2 and group 3 deletions, we have concluded that the two groE genes cannot be separated by more than 2,000 base pairs. We were unable to isolate any H18a deletion derivatives exhibiting the opposite spectrum of growth from that of group 3, that is, mutants that would grow on groEL- but not on groES­bacteria (group 4 deletion phages). Such deletions should only arise in phage H18a from a non-int gene-product-promoted recombination event originating to the left of or within the groE bacterial 45 genes. However, deletions with such properties should be generated by the gene product of int if the orientation of the bacterial piece of DNA relative to attA were reversed. We constructed such a phage derivative as follows. Phage groE+ vector 375 DNA (Fig. 1) was digested with EcoRI to release the two phage DNA arms and the groE+-containing bacterial DNA fragment. The three DNA fragments were religated and packaged in vitro, as described (7). Ten plaques were picked at random, phage stocks were grown, and their DNA restriction patterns were analyzed after BamHI digestion (21). Knowing that the EcoRI DNA fragment of I. coli carrying the groE+ genes contains an asymmetric BamHI restriction site, we anticipated that the BamHI DNA restriction pattern of phage carrying the groE+ genes in one orientation relative to the phage genes would be different from that of phage carrying the groE+ insert in the opposite orientation. Of the 10 phages, 6 were shown by BamHI digestion to nave the insert in one orientation (e.g., 375), whereas the other 4 had the opposite orientation (e.g., 378). We showed that phage H18a and 375 have the bacterial DNA in the same orientation by DNA heteroduplexing. This was verified by the isolation of group 3 phage deletions from phage 375, with plating properties identical to H18a613. Again, we were unable to isolate group 4 phage deletions. However, starting with phage 378, we were able to isolate group 4 phage deletions (i.e., those deletions exemplified by 3786101 and 3786 102) which plate only on groEL- but not on groES- bacterial mutants. As expected, we were unable to isolate group 3 phage deletion mutants from phage 378 because the relative orientations of 46 the groE bacterial genes and the attA site had changed. Identification of a second groE gene product To confirm the presence of the groES gene suggested by the genetic analysis, we examined the protein products made in UV­irradiated bacteria by the various AgroE+ transducing phage. Previously, we were only able to identify the 65,000 Mr protein as the groE gene product because the W3 groE- mutant changed its electrophoretic mobility (II). Phage deletion mutants of all classes were used to infect UV-irradiated bacteria, and the proteins made were labeled with 35s04-2, [35SJmethionine, or a mixture of 3H-amino acids and analyzed on NaDodS04/polyacrylamide gels. Group 1 and 4 groE deletion mutants induced the synthesis of the 65,000 Mr protein (Fig. 4). As expected, AgroE deletion mutants in groups 2 and 3 did not induce the synthesis of the 65,000 Mr protein. However, a protein band corresponding to an approximate Mr of 15,000 was present in AgroE+ and AgroE groups 1 and 3 deletion infections but was absent in AgroE groups 2 and 4 deletion infections (Fig. 4). The presence and absence of this protein species correlates exactly with the plating behavior of the various AgroE transducing derivatives and is consistent with its being the product of a second groE bacterial gene, which is mutated in all eleven groES- bacterial mutants (Table 2). In order to verify that the 15,000 Mr polypeptide is the product of the groES gene, we started with the A transducing phage H18a~13 and isolated mutants that were no longer able to form plaques 47 Figure 4. Autoradiograms of NaDodS04/polyacrylamide slab gel electrophoresis. (A) UV-irradiated 159uvrA-sup+ bacteria infected with various phage and labeled with a mixture of 3H-amino acids at 20 uCi/ml for 20 min at 37°C. Lanes: 1, H18a; 2, A~I60; 3, H18afi6; 4, H18afi13; and 5, H18afi3. The acrylamide concentration was a 10-25% linear gradient. (B) UV-irradiated 159 uvrA-sup+ bacteria infected with 378 (1), 378fi101, (2) and 378fi102 (3). The acrylamide was 12.5%. The arrows point to the positions of the 65,000 Mr and 15,000 Mr-groEL and groES polypeptides. 48 2 3 4 5 2 3 , , Table 2. Proteins induced by A groE+ transducing phages and their derivatives Phage strains* Relevant Protein, host Mr genotype A H18a 66t 63t 613t groEL~ml 378 65,000 sup+ + + + 65,000 supF +. + + + 15,000 sup+ + + + + + 15,000 supF + + + + + * +, Presence of the 65,000 Mr or 15,000 Mr protein species on NaDodS04/ polyacrylamide gels; - absence of the protein species. tPhages 66, 63, and 613, are deletion derivatives of H18a, whereas phage 6101 is a deletion derivative of 378. 6101t + + 01 a on groES bacteria. None of the - 100 independently isolated phage mutants behaved as expected for an amber nonsense mutation because none plated on groES30supE or supF bacteria. However, among 14 mutants tested, one, H18a~13 groESKT10, produced the 15,000 Mr protein at a normal rate after infection of UV-irradiated bacteria but altered the isoelectric point from pH 5.2 to 5.0 (Fig. 50). Revertants of H18a~13 groESKT10, isolated by their ability to grow on groES30 bacteria, were shown to recover simultaneously the wild-type isoelectric point (Fig. 5E). This result proves that the 15,000 Mr polypeptide is indeed the product of the groES gene. Discussion It has been demonstrated that transducing phages carrying the groE+ bacterial genes can be isolated as plaque formers on groE­mutant bacteria from a pool of transducing phages constructed in vitro so that they carry various segments of the I. coli chromosome (7 , 11,12) • All these A groE+ transduc i ng phages grow on all groE­bacterial mutants in our collection (11). However, the present studies indicate that our groE- bacterial mutants fall into two distinct complementation classes. GroEL- mutants, exemplified by groEL743 and groEL140, are mutated in the gene encoding the 65,000 Mr polypeptide, whereas groES- mutants, exemplified by groES7 and groES30, carry mutations in the gene encoding the 15,000 Mr polypeptide. The experimental basis for reaching this conclusion comprises the following data. (i) There exists a mutation in bacteriophage W3, called W3a, that affects both the ability to plate on groEL- bacteria and the electrophoretic mobility of the 65,000 Mr 51 Figure 5. Two-dimensional gel electrophoresis of infections of UV­irradiated 159uvrA-sup+ bacteria with H18a (A), H18~3 (B), H18a~13 (C), H18a~13groESKT10 (0), and H18a~13groESK10 revertant (E). Labeling was as described for Fig. 2B. Only the regions of the gel corresponding to a 4.75-5.75 pH gradient (in the horizontal dimension) and 12,000-17,000 Mr (in the vertical dimension) are shown. The solid arrows point to the position of the wild-type 15,000 Mr polypeptide. Open arrows point to the position of the H18a~ 13 groESKT10-coded 15,000 Mr polypeptide. 52 u lLJ CD polypeptide. W3a grows normally on groES- bacterial mutants. (ii) An amber mutation affecting the 65,000 Mr polypeptide similarly affects growth on groEL- but not on groES- bacterial mutants. (iii) Deletion derivatives of the H18a, 375, and 378 AgroE+ transducing phages exist that exhibit only one of the two groE complementing activities. Phages H18a and 375 have the bacterial DNA in the same orientation, with the relative order of the neighboring loci being (J+groES+groEL+attA). Phage 378, however, has the opposite order (i+groEL+groES+attA). Deletions of phages H18a and 375, starting at at-0 +, produce groES+ groELll but not groE~ groEL + deri vat i ves, whereas deletions of phage 378 produce groEsagroEL+ derivatives but not groES+groELll mutants. The existence of such deletions eliminates the possibility that the two observed groE complementation groups represent intragenic rather than intergenic complementation. (iv) There exists a perfect correlation between the presence of the 15,000 Mr polypeptide and groES+ complementation activity. Similarly, the presence of the 65,000 Mr polypeptide is always associated with groEL+ complementation activity. In spite of our efforts, we have been unable thus far to isolate amber mutations in the groES gene to further substantiate this conclusion. However, we have been able to isolate a transducing phage with a missense mutation of the groES gene causing a shift in the isoelectric point of the 15,000 Mr polypeptide from pH 5.2 to 5.0 (Fig. 5). Revertants of this mutant phage, isolated on the basis of their ability to form plaques on groES- hosts, make a 15,000 Mr protein with the wild-type isoelectric point. 54 · 55 We have been unable to detect any differences in the GroE phenotypes exhibited by mutants in the two groE genes. Mutations in either gene seem to affect A phage head morphogenesis at a very early step. The proheads isolated from A phage infections of groEL- or groES- bacterial mutants show the same molecular composition (i.e., gene products of I, Nu3, and l, and a small amount of unprocessed product!) (Fig. 2A). Furthermore, there are groE- mutants in either gene that are temperature sensitive for bacterial growth at 42°C. From these results it appears that the 65,000 Mr and 15,000 Mr gene products either act at the same levels in A phage and bacterial morphogenesis or act at separate steps of the same pathways. At this stage we cannot distinguish between these two possibilities. The position of groEL product action in A head morphogenesis has been established (7), yet purified preparations of the native decatetramers do not contain detectable amounts of the 15,000 Mr polypeptide, suggesting that the two groE proteins do not form a tight enough complex to copurify. In preliminary experiments, however, we have shown that mutations in the groEL gene can partially suppress mutations in the groES gene, suggesting that the proteins do indeed interact in vivo (unpublished results). Elucidating the mode of action of the two groE bacterial gene products in bacteriophage A, T4, and T5 morphogenesis and in I. coli growth will contribute significantly to our understanding of the assembly mechanism of complex macromolecular structure. Literature Cited 1. Georgopoulos, C.P. & Eisen, H (1974) J. Supramo1. Struct.!, 349-359. 2. Georgopoulos, C.P., Hendrix, R.W. Kaiser, A.D. & Wood, W.B. (1972) Nature (london) New Bio1. 239, 38-41. 3. Zweig, M. & Cummings, D.J. (1973) J. Mol. Bio1. 80, 505-518. 4. Georgopoulos, C.P., Hendrix, R.W., Casjens, S. & Kaiser, A.D. (1973) J. Mol. Bio1. ~, 45-60. 5. Sternberg, N. (1973) J. Mol. Bio1. ~, 25-44. 6. Hendrix, R.W. & Casjens, S. (1975) J. Mol. Bio1.~, 187-199. 7. Muria1do, H. (1979) Viro10gy~, 341-367. 8. Takano, T. & Kakefuda, T. (1972) Nature (london) New Bio1. 239, 34-37. 9. Coppa, A., Manzi, A., Pulitzer, J.F. & Takahashi, H. (1971) J. Mol. Bio1. ~, 61-87. 10. Revel, H.R., Stitt, B.l., Lie1ausis, I. & Wood, W.B. (1980) J. Vi ro1. ~, 366-376. 11. Georgopoulos, C.P. & Hahn, B. (1978) Proc. Nat1. Acad. Sci. USA ]2, 131-135. 12. Hendrix, R.W. & Tsui, L. (1978) Proc. Nat1. Acad. Sci. USA~, 136-139. 13. Hendrix, R.W. (1979) J. Mol. Bio1. 129,375-392. 14. Hahn, T., Hahn, B., Engel, A., Wurtz, M. & Smith, P.R. (1979) J. Mol. Bio1. 129,359-375. 15. Murray, N.E., Brammar, W.G. & Murray, K. (1977) Mol. Gen. Genet 150, 53-61. 16. Thomas, M., Cameron, J.R. & Davis, R.W. (1974) Proc. Nat1. Acad. Sci. USA 11, 4579-4583. 17. Parkinson, J.S. & Huskey, R. (1971) J. Mol. Bio1.~, 369-384. 18. Fowler, R.G., Degnen, G.E. & Cox, E.C. (1974) Mol. Gen. Genet. 133, 179-191. 19. Carroll, D., Ajioka, R.S. & Georgopoulos, C.P. (1980) Gene.lQ., 261-272. 56 20. O'Farrell, P.H. (1975) J. Bi01. Chern. 250,4007-4021. 21. Enquist, L.W., Madden, M.J., Shiop-Stans1y, P. & Vande Woude, G.F. (1979) Science 203, 541-544. 57 JOURNAL OF BACTERIOLOGY, Mar. 1982, p. 1082-1088 0021-9193/82/031082-07$02.00/0 Vol. 149. No.3 Evidence That the Two .Escherichia coli groE Morphogenetic Gene Products Interact In Vivo KIT TILLY* AND COSTA GEORGOPOULOS Department of Cellular, Viral, and Molecular Biology, University of Utah Medical Center, Salt Lake City, Utah 84132 Received 23 June 1981/Accepted 12 October 1981 The Escherichia coli groEL and groES gene products are essential for both phage morphogenesis and bacterial growth. Although the gene products have been identified, their exact roles in these processes are not known. We have isolated mutations in the groEL gene that suppress defects in the groES gene. These intergenic suppressors were shown to map in the groEL gene by a variety of genetic and biochemical analyses. These results suggest that the two morphoge­netic gene products interact in vivo and help to explain why mutations in either gene exhibit the same phenotype with respect to A head assembly and bacterial growth. Although Escherichia coli groE mutants were isolated as hosts in which bacteriophage A is unable to undergo productive infection, such mutants were subsequently shown to be pleio­tropic, also affecting the growth of phages T4 and T5 and the bacterial host itself (2, 3, 12, 13), All of the effects on phage growth seem to be at the level of morphogenesis. Bacteriophages A and T4 are unable to make functional heads in groE hosts, whereas T5 makes heads normally but only undergoes abortive tail assembly (16). The groE gene products must be necessary for bacterial growth, since many groE mutants are temperature sensitive for bacterial growth, forming long filaments without septa at 42°C (1). A groE+ transducing phage have been isolated from pools of EcoRI and HindIlI fragments of E. coli DNA ligated into various phage vectors (4, 6, 8). Such phage are able to form plaques on all known groE bacterial mutants, indicating that they carry all groE genes in their bacterial DNA (4). Analysis of point and deletion mutations of these phage has led to the identification of two groE genes and their products (4, 6, 15). One gene product (designated gpgroEL, since it is the larger groE gene product) is a protein of 65,000 Mr which has been purified to homogeneity (5, 7). In its native state gpgroEL is a decatetramer, with its 14 subunits arranged in a double ring with sevenfold symmetry. This complex sedi­ments at 25S and has a weak ATPase activity (5, 8). gpgroES has recently been shown to be a protein of 15,000 Mr (15). Genetic studies have helped to clarify the position at which the groE gene products act in phage morphogenesis. A mutants able to over­come the groE block (called At) arise at a frequency of 10-7 to 10-8 and have mutations in either the B or the E gene (2). The A E gene mutants seem to lower the levels of active gpE in the infected cell, allowing the residual activity of a mutant gpgroE to complete a few A heads. Murialdo (8) has shown by sedimentation ex­periments that at least gpgroEL interacts with gpB at a very early stage in A head assembly. Out of 11 groEL mutants in our collection, only 1, mutant groE1A4, does not allow T4 growth. In addition, we found that six out of six hd bacterial mutants of Revel et al. (11), which block T4 head assembly at the level of gp3J action, map in the groEL gene (unpublished results). These findings suggest that gpgroES may be nonessential for T4 head assembly. The identification of a second groE gene raises the possibility of cooperation between the two groE gene products. The A proheads formed upon infection of bacteria with mutations in either groE gene have a similar composition, suggesting that gpgroES may act at the same step as gpgroEL in A head morphogenesis (15). In this paper we describe genetic and biochemi­cal experiments indicating that the proteins in­teract in vivo, suggesting that the two proteins act concomitantly, rather than sequentially. MATERIALS AND METHODS Phage and bacterial strains. E. coli groE mutants derived from B178, all of which have the same pheno­type with respect to A growth, have been shown to fall into two complementation groups (15). The groES619 strain is a representative of one of these groups, carrying a mutation in the gene encoding the 15,()()().Mr groE polypeptide which prevents normal A head as­sembly and makes the bacterium temperature sensi­tive for growth. T4 grows norma1ly in strain groES619 and all other groES mutants. The groELA4 strain has a VOL. 149, 1982 mutation in the groE gene encoding the 65,000-Mr protein, blocks A head morphogenesis at the same step as strain groES619, and is also temperature sensitive for bacterial growth. T4 does not make functional heads in groEIA4 bacteria (3). The construction of A groE+ transducing phage carrying either or both of the groE genes has been previously described (4, 8, 15). T4E mutants are T4 derivatives able to bypass the groE block found in a specific groEL mutant [i.e., T4E1 was selected for growth on strain groEL44, whereas T4E3 was selected for growth on strain groES619(Ts+)4]. Isolation of temperature-resistant revertants. Tem­perature- resistant revertants of groES619 were isolat~ ed by incubating groES619 bacteria on T plates at 4rC for 2 days and were obtained at a frequency of about 10-7 to 10-6 • Labeling experiments. Bacteria to be labeled were grown overnight at 30°C in high-sulfur M9 medium (4) supplemented with all amino acids except methionine and cysteine. They were diluted 1110 into the same medium, grown until the cultures had reached a con­centration of 3 x 108 bacteria per ml, and shifted to 42°C for 10 min. They were then labeled with 20 !-LCi of [l5S]methionine per ml (1,380 Cilmmol, Amersham S1204) for 10 min at 4rC. The labeling was stopped by centrifugation, and the pellets were resuspended in 0.05 volume of two-dimensional lysis buffer (10). T4 proteins were labeled by infecting log-phase (3 x 108/ mI) cultures of the bacteria of interest with T4 at a multiplicity of infection (MOl) ofl0, growing 20 min at 37°C with aeration, and labeling with 80 !-LCi of [35S]methionine per mI for 10 min. The cultures were centrifuged for 1 min at 15,000 x g, and the pellets were resuspended in 0.20 volume of sodium dodecyl sulfate sample buffer. One- and two-dimensional polyacrylamide gel electro­pboresis. These were as previously described (15). Complementation test with T4 phage mutants. About 2 x lOS groEL140 sup+ bacteria were mixed with 108 T4 amNG71 (or another T4 amber or E mutant), 3 mlof soft agar was added, and the mixture was plated on L plates. Approximately 20 !-LI of various phage suspen­sions (containing about 3 x 106 phage) was placed on top of the soft agar and allowed to dry. After overnight incubation at 37°C, positive complementation resulted in the confluent lysis of the bacterial lawn at the site of the drop, whereas negative complementation resulted in few or no plaques. The results of spot complementa­tion tests were occasionally verified by measuring phage yield after coinfection of liquid cultures of groEL140 sup+ bacteria with an MOl of 10 of each T4 mutant strain. Media and bacterial and pbage platings. These were as previously described (2, 4). Phage yield experiments. The effect of preinfection by various A groE transducing phage on T4 infection was determined by phage yield experiments. The bacteria to be tested were grown to 3 x lO8/ml, and various A phage were added at an MOl of 10. After 20 min of adsorption at room temperature, the cultures were shaken at 37°C for 30 min. T4 or T4E1 was then added at an MOl of 5, and the cultures were shaken at 37°C for 10 min. The cultures were spun for 1 min at 15,000 x g, and the supernatants (containing unad­sorbed phage) were discarded. The pellets were resus­pended in 0.1 ml of L broth and shaken another 50 min at 37°C. A few drops of CHCh were added to complete E. COLI groE GENE PRODUCTS lysis, and the T4 progeny were assayed on E. coli B, on which host phage A does not form plaques. RESULTS Isolation of suppressors of groES619. Experi­ments were designed to determine whether the apparent concerted action of the groE gene products involved protein-protein interactions between them. We reasoned that if we could suppress mutations in one gene by mutating the second gene, this would support this hypothesis. The only groE mutants shown to block T4 growth (some of which permit A growth [11]) map in the groEL gene (unpublished data), so this distinguishing phenotype was used for se­lecting altered groEL genes. The strategy used for isolating specific intergenic suppressors was to start with groES mutants which are tempera­ture sensitive for bacterial growth, isolate tem­perature- resistant derivatives, and screen those survivors for inability to propagate phage T4 or T4E1 (a T4 mutant able to grow on strain groEIA4). Both T4 and T4El propagated normal­lyon all groES mutants tested. Several classes of temperature-resistant bac­teria were obtained. Strain groES619(Ts+)1 is a representative of the first and most populous class, isolated at a frequency of 5 x 10-7 • These bacteria now permit the growth of all A and T 4 derivatives (Table 1), and they probably repre­sent true revertants of the groES619 mutation. Strains groES619(Ts+)7 and (Ts+)10 are repre­sentatives of the second class, found about 20- fold less frequently than the revertants in the first class. On these hosts, A grows normally, T4 makes small plaques at a close to wild-type efficiency, and T4E1 forms plaques at a frequen­cy of less than 10-4 (Table 1). The third class [e.g., groES619(Ts+)4] does not allow the growth of T 4 or T 4E 1, while permitting A to grow normally. Since the last two classes of tempera­ture- resistant revertants have the phenotype of groEL mutants, they were subjected to further analysis. The second and third types of temperature­resistant revertants (putative groEL mutants) were isolated from strains groES619 and groES42 at a frequency of 1 to 5% of all tempera­ture- resistant revertants, but not from strain groES7 (although at least 1,000 temperature­resistant revertants were tested). The frequen­cies with which temperature-resistant bacteria were obtained in all cases were comparable, so the mutations required to give such a phenotype must be allele dependent. Isolation of T4E mutants on temperature-resis­tant hosts. When T4 phage were plated on the groES619(Ts+) mutants that were not true re­vertants, plaques were formed at a frequency of 59 TILLY AND GEORGOPOULOS J. BACTERIOL. TABLE 1. Plating properties of various phage on groE strains Efficiency of plating" of phage strain: Bacterial strain A groES+ A groES+ A ~(groES) A groEL+ ~(groEL) groEL+ T4 T4d T4e3 B178 groES+ groEL + + + + + + + + groEL44 + + + groEL140 + + + + groES619(Ts) + + + + + groES619(Ts +)1 + + + + + + + groES619(Ts +)4 + + + + + groES619(Ts+)4 (A imm21 groES+ groEL +) + + + + + + + groES619(Ts +)7 + + + + ± + groES619(Ts +)7 (A imm21 groES+ groEL +) + + + + + + + groES619(Ts +)10 + + + + ± + groES619(Ts+)10 (A imm21 groES+ groEL +) + + + + + + + a (+) Denotes large plaque size and an efficiency of plating of 0.5 to 1.0 at 37°C when the efficiency of plating on B178 is taken to be 1.0. (±) Denotes an efficiency of plating of 0.1 to 1.0 with small plaques, and (-) denotes an efficiency of plating of less than 10-4 • approximately 10-6• One of these [T4E3, isolat­ed on strain groES619(Ts +)4] was grown up and tested on various groE and temperature-resis­tant revertant strains (Table 1). Since T4E3 was found to be unable to grow on groEL140 bacte­ria, on which T4 wild type can grow, this bacte­rial host was used for mapping the position of the s mutation. In complementation tests on groEL140 sup+ bacteria, T4s3 was able to com­plement phage bearing amber mutations in all genes tested except gene 31, placing the muta­tion in that gene. The result shows that the inability of phage T4s3 to propagate on strain groEL140 is at the level of gp31 action. Since a mutation in gene 31 can overcome the block on T4 growth found in groES619(Ts+)4 bacteria, it is likely that T4 cannot grow because wild-type gpl1 is inactive in this host. Phenotype of T4 infection. T4 infection of groEIA4 bacteria is blocked at the level of gp31 action, resulting in the absence of the proteolytic cleavages found in normal head assembly (3) (Fig. 1, lane 13). The protein profiles after T4 infection of groES619(Ts+) derivatives were compared with those in strains groEIA4 and groES619 to determine the position of the block in T4 growth. The proteins synthesized in groEIA4, groES619, and two temperature-resis­tant revertants [(Ts +)4 and (Ts +)7] were labeled between 20 and 30 min after infection with T4 (or its derivatives T4E1, T4E3, orT431- [amNG71]). Sodium dodecyl sulfate-polyacrylamide gel elec­trophoresis of the proteins revealed that the pattern of gp23 cleavage exactly parallels the pattern of phage growth; gp23 (and two other proteins, which have not been identified) is not cleaved when the phage-host combination is nonproductive (Fig. 1). There are no differences between the protein profiles found in T4-infect-ed groEIA4 bacteria and T4 amNG71-infected groES619 bacteria, in which cases the blocks have been proven to be at the level of gp31 function, and those found after T4 infection of groES619(Ts +) derivatives. These results sug­gest that strains groES619(Ts+)4 and (Ts+)7 block T4 and T4E1 phage growth at the level of action of gp31 . Lysogenization experiments. The third step in analyzing the groES619(Ts +) revertants was to lysogenize them with A imm21 cl + transducing phage bearing both of the groE genes. The results of the lysogenization experiments are presented in Table 1. In all cases, lysogenization with a A imm21 cl+ groES+ groEL + transducing phage made the strain wild type with respect to both A and T4 growth. This is because the wild­type genotype is dominant over the mutant alleles, and the prophage copies of the groE genes are expressed at the same levels as the bacterial copies (see Fig. 2) (4). T4 phage yield experiments. We determined T4 phage yield in various bacterial strains both with and without pre infection by various A groE transducing phage. Similar results were obtained when groES619(Ts+)4 bacteria were infected with T4 or groES619(Ts+)10 bacteria were in­fected with T4E1 (Table 2). In these cases, the yield of T4 or T4E1 was reduced by a factor of 100 to 500 from that found in either B178 groES+ groEL + or groES619(Ts). Preinfec­tion with A imm21 cl+ or A imm21 cl+ groES+ il(groEL) for 30 min had no effect on the T4 or T4E1 phage yield. Preinfection with either A imm21 cl+ groES+ groEL + or A imm21 cl+ il (groES) groEL +, however, had the effect of increasing the T4 or T4E1 phage yield by a factor of 100 to 200. In this series of experiments, the presence of wild-type gpgroEL suppressed the 60 VOL. 149, 1982 E. COLl groE GENE PRODUCTS ~ .... ,~~ - ~ .... ......... ..- ... -*' _ .. ' ... ta. 2 3 4 5 6 7 B 9 10 II 12 13 14 15 16 FIG. 1. Autoradiograms of sodium dodecyl sulfate-polyacrylamide slab gel electrophoresis. Various hosts were infected with phage T4 (lanes 1,5,9, 13), T4E1 (lanes 2, 6, 10, 14), T4E3 (lanes, 3, 7, 11, 15), or T4 amNG71 (31 - ) (lanes 4, 8, 12, 16) and labeled with eSS]methionine at 40 fLCi/ml from 20 to 30 min after infection. The bacteria were groES619 (lanes 1 to 4), groES619(Ts+)4 (lanes 5 to 8), groES619(Ts+)7 (lanes 9 to 12), and groEIA4 (lanes 13 to 16). The arrows mark the positions of two unidentified proteins that are only seen when gp23 is cleaved. The acrylamide was 12.5%. block on T4 or T4el growth, and supplying A proteins alone or gpgroES plus A proteins was not sufficient to compensate. Biochemical analysis. Because the phenotype of the revertants, the lysogenization experi­ments, and the phage yield results all pointed to the groEL gene as the site of the suppressor mutations, some of these mutations might be expected to alter the characteristics of the groEL polypeptide. Since gpgroEL is known to be synthesized at a relatively higher rate after a shift from 30 to 42°C (9; D. Drahos and R. W. Hendrix, J. Bacteriol., in press; unpublished data), bacteria were labeled with [35S]methio­nine between 10 and 20 min after such a shift to increase its molar amount in the cell. The bacte­rial proteins made during such a pulse were displayed on two-dimensional polyacrylamide gels. Strain groES619(Ts+)7 and 10 other strains were shown to encode gpgroEL with a wild-type isoelectric point (Fig. 2b). However, strain groES619(Ts+)4 and two other strains were shown to possess gpgroEL with a more basic isoelectric point (Fig. 2c). The magnitUde and direction of this shift are most apparent when (Ts+)4 bacteria lysogenic for A groES+ groEL + are heat shocked and labeled (Fig. 2d). In this case, two spots replace the normal gpgroEL spot. Because the intensities of the spots are similar, the wild-type protein encoded by the A transducing phage and the mutant protein from the bacterial chromosome appear to be synthe­sized in approximately equal amounts. Since such a lysogen regains the wild-type phenotype (Table 2), the mutation must be recessive. The discovery of a spontaneous change in the bio­chemical properties of gpgroEL coincident with the genetic results confirm our hypothesis that a mutation in the groEL gene can suppress the phenotype of a mutation in the groES gene. DISCUSSION The evidence that the groEL and groES gene products interact stems from the isolation of bacteria in which a groEL mutation has sup­pressed a groES mutation. The suppressor mu­tations, which make strain groES619 tempera­ture resistant for growth at 42°C, are in the groEL gene by four criteria. (i) The phenotype of the temperature-resistant revertants is that of groEL mutants. T4el and, in some cases, T4 do not grow, and the proteins normally cleaved during head morphogenesis of these phage re­main in their uncleaved forms. T4E3, which is a T4 mutant able to grow on all the groES619(Ts +) mutants, has an altered gene 31 protein, showing that the block in T4 infection is at the level of gp3/ action, as is found in known groEL mu­tants. (ii) Lysogenization with a A imm21 groES+ groEL + phage is sufficient to restore the ability of T4 to grow on temperature-resistant rever- 61 TILL Y ANO GEORGOPOULOS • - - - • c - - -• , .... ...~ "... J. BACTERIOL. b - d FIG. 2. Two-dimensional gel electrophoresis of bacterial proteins labeled with 20 fLCi of [l5S]methionine per mI from 10 to 20 min after a shift from 30 to 42°C. The bacteria were (a) groES619, (b) groES619(Ts+)7, (c) groES619(Ts+)4. and (d) groES619(Ts+)4 (JI. imm21 groES+ groEL +). Only the regions of the gels corresponding to a 6.0 to 4.5 pH gradient (in the horizontal dimension) and approximately 58,000 to 75,000 molecular weight (in the vertical dimension) are shown. The arrow points to the position of wild-type gpgroEL. TABLE 2. T4 phage yield in groES619(Ts+) bacteriao Bacterial host Preinfecting phage B178 groES+ groEL + none lI. imm21 cI+ lI. imm21 cI+ groES+ groEL + lI. imm21 cI+ groES+ A(groEL) lI. imm21 cI+ A(groES) groEL + groES619(Ts+)7 none lI. imm21 cI+ lI. imm21 cI+ groES+ groEL + lI. imm21 cI+ groES+ t:.(groEL) lI. imm21 cI+ t:.(groES) groEL + groES619(Ts +)4 none lI. imm21 cI+ lI. imm21 cI+ groES+ groEL + lI. imm21 cI+ groES+ t:.(groEL) lI. imm21 cI+ t:.(groES) groEL + groES619(Ts+)J 1 none lI. imm21 cI+ lI. imm21 cI+ groES+ groEL + lI. imm21 cI+ groES+ A(groEL) lI. imm21 cI+ t:.(groES) groEL + ° The phage yield experiments were done as described in the text. b NO, Not determined. Phage yield T4 T4€1 99.5 51 122.5 148 74 59 85.5 83.5 64 92 3.4 NOb 3.6 NO 37.2 NO 1.5 NO 58.2 NO 0.2 NO 0.2 NO 37.2 NO 0.2 NO 35.4 NO NO 0.3 NO 0.4 NO 29.0 NO 0.3 NO 27.4 62 VOL. 149, 1982 tants, whereas lysogenization with A imm21 alone is not sufficient. (iii) The reduced T4 phage yield found in temperature-resistant revertants was increased by a factor of 100 to 200 by preinfection with A imm21 groES+ groEL + or A imm21 il(groES) groEL +, whereas preinfection with A imm21 or A imm21groES+ il(groEL) had no effect. (iv) Of 14 temperature-resistant rever­tants, 3 synthesized groEL polypeptide with an altered isoelectric point. The combination of genetic behavior with the presence of a physical­ly altered groEL protein in mutants obtained without mutagenesis places the suppressor mu­tations in the groEL gene. Our ability to isolate such suppressors is strong evidence that the groE gene products interact in vivo. The existence of interaction between gpgroEL and gpgroES suggests the possibility of the two proteins acting together at the same step. Action at the same step is consistent with our observa­tions that th~ abnormal A proheads found in mutants in either gene are identical (15). There are mutants in both genes that block T5 tail assembly, and there are mutants in both genes that are temperature sensitive for bacterial growth. We have not obtained any evidence that the groES protein is required for T4 morphogen­esis. There may be more than one active site on a gpgroEL-gpgroES complex, one of which is composed of parts of both proteins (and is required for A, T5, and E. coli growth) and one of which consists of only groEL protein (and is required for A, T4, T5, and E. coli growth). Changes in either site may not be independent, because the mutants we obtained must affect both. Alternatively, there may be only one ac­tive site, but the levels of functional groE com­plex required for each system are different. It is likely that we have not studied all groEL suppressors of the groES619(Ts) mutation, since we only analyzed the 1 to 5% of the temperature­resistant survivors that blocked T4 or T4e1 growth. It is clear that the groEL gene can be mutated without affecting T4 growth, since only 1/11 of the groEL mutants in our collection (none of which allow A morphogenesis) block T4 head assembly. Therefore, many more of the survivors that we obtained may carry groEL suppressors. In fact, at least three independent temperature-resistant revertants had the same shift in isoelectric point, indicating that we may be analyzing a few specific mutations that are capable of compensating for the groES619 muta­tion. An important observation was that we were unable to isolate temperature-resistant rever­tants that block T4 growth from some groES mutants. These strains may either have muta­tions that cannot be compensated for by a viable groEL mutation (e.g., affecting part of gpgroES E. COLI groE GENE PRODUCTS that does not interact with gpgroEL) or muta­tions that can be suppressed by groEL mutation without affecting the protein's activity in T4 head assembly. The existence of such allele specificity for suppression in groES mutants strongly argues against the possibility that two independent mutations cause the temperature­resistant and T4 nonpermissive phenotypes. In addition, spontaneous groES619(Ts +) rever­tants which simultaneously block T4 head mor­phogenesis represent 1 to 5% of the surveyed population. This frequency is much higher than that at which one would expect to find double mutants in the absence of mutagenesis. Although these studies suggest that the groE proteins act together as a complex in bacterial growth and phage morphogenesis, no biochemi­cal evidence for the existence of such a complex has yet been obtained. The groES protein does not sediment with the 255 groEL protein decate­tramer on glycerol gradients nor does it precipi­tate with anti-gpgroEL antibodies (unpublished observations). Further analysis of the nature of the interaction between the two proteins and their function in bacterial and phage morphogen­esis must await the fine-structure mapping of specific mutations and in vitro studies on the purified proteins. ACKNOWLEDGMENTS We thank Timothy Hey for his contributions at the initial stages of this project, Glenn Herrick for constructive criticism of the manuscript, and Jerri Cohenour for typing the manu­script. This work was supported by Public Health Service grants GM23917 and GM07464 from the National Institutes of Health. LITERATURE CITED 1. Georgopoulos, C. P., and H. Eisen. 1974. Bacterial mu­tants which block phage assembly. J. Supramol. Struct. 2:349-359. 2. Georgopoulos, C. P., R. W. Hendrix, S. Casjens, and A. D. Kaiser. 1973. Host participation in bacteriophage lambda head assembly. J. Mol. BioI. 76:45-60. 3. 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. Nat. New BioI. 239:38-41. 4. Georgopoulos, C. P., and B. Hobn. 1978. Identification of a host protein necessary for bacteriophage morphogenesis (the groEproduct). Proc. Natl. Acad. Sci. U.S.A. 75:131- 135. 5. Hendrix, R. W. 1979. Purification and properties of groE, a host protein involved in bacteriophage assembly. J. Mol. BioI. 129:375-392. 6. Hendrix, R. W., and L. Tsui. 1978. Role of the host in virus assembly: cloning of the Escherichia coli groE gene and identification of its protein product. Proc. Natl. Acad. Sci. U.S.A. 75:136-139. 7. Hobn, T., B. Hobn, A. Engel, M. Wurtz, and P. R. Smith. 1979. Isolation and characterization of the host protein groE involved in bacteriophage lambda assembly. J. Mol. BioI. 129:359-373. 8. Murialdo, H. 1979. Early intermediates in bacteriophage lambda prohead assembly. Virology 96:341-367. 9. Neidhardt, F. C., T. A. Phillips, R. A. VanBogelen, M. W. 63 TILLY AND GEORGOPOULOS Smitb, Y. Georgalis, and A. R. Subramanian. 1981. Identi­ty of the B56.5 protein. the A-protein. and the groE gene product of Escherichia coli. J. BacterioL 145:513-520. 10. O'Farrell, P. H. 1975. High resolution two-dimensional electrophoresis of proteins. J. BioI. Chem. 250:4007-4021. 11. Revel, H. R., B. L. Stitt, I. Lielausis, and W. B. Wood. 1980. Role of the host cell in bacteriophage T4 develop­ment I. Characterization of host mutants that block T4 head assembly. J. Virol. 33:366-376. 12. Sternberg, N. 1973. Properties of a mutant of Escherichia coli defective in bacteriophage A head formation (groE). I. Initial characterization. J. Mol. BioL 76:1-24. 13. Sternberg, N. 1973. Properties of a mutant of Escherichia J. BACTERIOL. coli defective in bacteriophage A head formation (groE). II. The propagation of phage A. J. Mol. BioL 76:25-44. 14. Takano, T., and T. Kakefuda. 1972. Involvement of a bacterial factor in morphogenesis of bacteriophage cap­sid. Nat. New BioI. 239:34-37. 15. Tilly, K., H. Murialdo, and C. Georgopoulos. 1981. Identi­fication of a second Escherichia coli groE gene whose product is necessary for bacteriophage morphogenesis. Proc. Natl. Acad. Sci. U.S.A. 78:1629-1633. 16. Zweig, M., and D. J. Cummings. 1973. Cleavage of head and tail proteins during bacteriophage T5 assembly: selec­tive host involvement in the cleavage of a tail protein. J. Mol. BioI. 80:505-518. 64 APPENDIX Further Analysis of a groES619 Temperature-resistant Revertant An important confirmation that temperature-resistance results from a single mutational event, rather than reversion of the groES­mutation and simultaneous mutation of the groEL gene, was the separation of the two mutant alleles. The groES619 and groEL4 (from roES619 ~+4) gene products can be distinguished from the wild-type because each mutation results in a basic shift of the isoelectric point of the affected protein. These phenotypes were used to follow the segregation of the alleles. The strategy used to determi ne the ~ genotype of the roES619ts+4 strain was to transfer the chromosomal alleles onto transducing phage and to analyze the plating behavior and protein profiles of those phage. AgroES+groEL+n4, which has a deletion from the A attachment (at0) site extendi ng past the i nt gene, was used to lysogenize groES619ts+4. The deletion of attA and int increased the probability that the phage would integrate by homologous recombi nati on into the bacter; a 1 genes. Ill. lysogen was induced with ultraviolet light (950 ergs/mm2) and stocks were grown of six of the resultant phage. The phage were tested for growth on various bacterial strains (Table 1) and used to infect UV-irradiated bacteria and program the synthesis of labeled phage proteins (Fig. 1). Three of the six phage behaved genet~cally as if they were still AgroES+groEL+ and their protein profiles confirmed their genotype (Fig. lA). One phage grew well on wild-type or groEL- bacteria but poorly on groES- strains. Two-dimensional gel analysis showed its genotype to be AgroES619groEL+ (Fig. 18). The final two phages formed small plaques on all groE strains. Their protein patterns showed them to be AgroES619groEL4 (Fig. Ie). The groEL4 variant was isolated in a separate but similar screen. A AgroESAgroEL+ transducing phage, which has a deleted attA site but retains the int gene, was used to lysogenize groES619ts+4 bacteri a. The prophages we re induced and tested as befo re. Several were isolated that no longer grew on groEL- bacteria. All of these were shown to encode the groEL4 isoelectric variant (data not shown). Three conclusions can be drawn from this series of experiments. First, the groES619 mutation is still present in the groES619ts+4 strain, showing that temperature-resistance is a consequence of suppression, rather than simultaneous reversion to ~roES+ and mutation of groEL+ to groEL4. Second, the groEL4 mutation in combination with groES+ is non-productive for A growth. Third, the groES619groEL4 combination on a A transducing phage allows that phage to grow on groES619 bacteria. These conclusions further support the idea that there is functional interaction between the two ~roE gene products. 66 TABLE I. Plating properties of various phage strains Bacterial Strain Phage Strain 1 2 3 4 5 6 AgroEL4 B178 groES+ groEL+ groES30 groEL+ groES+ 9roEL140 + + + + + + ± ± + - + + ± ± + + + + + + means an efficiency of plating (eop) of 1 (when growth on B178 is the standard ), ± means an eop of 1 with small plaques, and - means an eop of 10-2• Phage strains were from integration of AgroES+ groEL+~ 4 into the chromosomal groE genes of 9roES619tS+4, followed by ultra violet irradiation of the lysogens and excision of the prophage. 67 Figure 1. Correlation of groE protein isoelectric variants with genotypes. Two-dimensional gel electrophoresis of lysates of 159uvrA­bacteria infected with A~4-3 (A), A~4-4 (B), and A~4-1 (C). Infections were as described in the chapter. The horizontal dimension, isoelectric focusing, was from basic (left) to acidic (right). The vertical dimension was NaDodS04-12.5% acrylamide slab gel electrophoresis. The arrows indicate the positions of the wild­type groE gene products (65,000 Mr and 15,000 Mr ). 68 CHAPTER IV GROE GENE REGULATION Introduction The I. coli groEL and groES genes were discovered because mutants block productive ~ infection at the level of head assembly (1,2). The genes were shown to be essential for bacterial growth, since mutants in either gene are temperature-sensitive, forming long filaments without septa at 42°C (3). Both genes are also required for proper T5 tail assembly (4). So far, however, only active groEL gene product has been shown to be required for T4 head assembly (5; K.Tilly and C. Georgopoulos, unpublished results). Much work has concentrated on gpgroEL and its function in head morphogenesis. The groEL gene product has been identified as a protein of 65,000 Mr (6,7). It has been purified to essential homogeneity and has a weak ATPase activity in vitro (8,9). In its native state, it forms a decatetramer with seven-fold symmetry, which sediments at 25S in velocity gradients (8,9). The oligomer seems to interact with the ~ ! protein in the very early stages of head assemb ly (1,10). The groES gene product is a protein of 15,000 Mr. It is either only loosely associated with the gp!-gpgroEL complex or completely absent (K. Tilly and C. Georgopoulos, unpublished results). The phenotypes of A infection in groEL- and groES­bacteria are identical (11), so gpgroES probably also acts early in A assembly. Little is known about the functions of the groE gene products in bacterial growth. Early studies focused on the phenotypes of temperature-sensitive groE- mutants at the restrictive temperature (3,12). The bacteria have defective septum formation and altered protein synthesis patterns (3). One groEL- mutation has been shown to also affect the levels of ~RNA and J:RNA at the non-permissive temperature, although this phenotype seems to require an additional, unlinked mutation (12). More recent experiments have shown that synthesis of the groEL gene product is induced by many stresses on the cell, providing another approach to the problem of groE protein function. The best characterized stimulus inducing gpgroEL synthesis is heat-shock (13,14). The heat-shock response in I. coli is characterized by a transient increase in the relative rates of synthesis of at least eight proteins, with most other proteins continuing to be synthesized at their normal rates (15,16). The eight genes seem to be controlled by a single positive regulatory locus, named htpR (13) or hin (14). A bacterial strain with an amber mutation in the htpR gene and a temperature-sensitive ~ allele has been isolated (17). This strain fails to induce the synthesis of the heat-shock polypeptides at 42°C and the cells lyse after prolonged exposure to the high temperature (13). This mutant is also unprotected from thermal 71 killing at 54°C by pregrowth at 42°C (14). Three of the heat-shock polypeptides have been identified as gpgroEL, gpdnaK, and form II of lysyl-tRNA synthetase (13,14,18). Yamamori and Yura (14) found that heat-shock led to a seven-fold increase in the amount of RNA that will hybridize in liquid to a AgroE+ transducing phage, suggesting that induction is controlled at the level of transcription. Lowered levels of the RNA polymerase subunit (j (19) and infection with A (20,21) also seem to induce the synthesis of all of the heat-shock polypeptides. The different inducing stimuli probably exert their effects via the ~ pathway (19; C. Georgopoulos, u'npublished results). In addition, the synthesis of at least gpgroEL is induced by amino acid starvation (22). The mechanism of this induction is unknown. This paper will describe the analysis of groE5 gene control. The groE5 gene seems to be co-~ranscribed and co-regulated with the .groEL gene, suggest i ng that it too is contro 11 ed by the htpR gene product. Methods and Materials ~acterial and phage strains The groE- bacterial strains and AgroE transducing phages were described previously (2,23; and ref. therein). K803 is hsdR- ~ -.?upF. Labeling experiments Bacterial and phage proteins were labeled with [355J 72 methionine as described previously (2,23) with bacterial labeling from 5 to 15 min. after the temperature shift to 42°C. Restriction enzyme digests and agarose gel electrophoresis All restriction enzymes were purchased from New England Biolabs or Bethesda Research laboratories and digestion conditions were those recommended by the supplier. Sau3A partial digestions were done as follows: a) 0.35u of Sau3A were added to 2 ug of DNA in the specified reaction conditions, b) equal aliquots were removed at 6,12,18,24 and 30 min and the reaction was stopped with EDTA and c) the pooled portions were ligated with BamHI - and bacterial alkaline phosphatase-digested pJB8. Restriction fragments were separated on 1% agarose gels (Seakem, Marine Colloids) in E buffer (40 mM Tris (pH 7.8), 20 mM Na acetate, and 2 mM EDTA). ligation and transformation Restriction enzyme digests were ligated with T4 ligase as recommended by the supplier (P.l. Biochemicals). ligation reactions were phenol-extracted and precipitated with isopropanol before transformation (which was as described by Mandel and Higa (24)). Large-scale plasmid preparations for transformation were as described by Clewell and Helinski (25), except that DNA was precipitated with isopropanol and equilibrium centrifugation was in Beckman 50.2 Ti rotors (31,000 rpm for approximately 40 hr). RNA and DNA transfer and hybridization RNA (prepared by the method of Gegenheimer and Apirion (26)) 73 was run on methyl-mercury hydroxide - 1.2% agarose gels, transferred onto diazobenzyloxymethyl paper, and hybridized with nick-translated (27) DNA probes (28; 107 to .108 dpm/ug; using 32p -dCTP from Amersham). DNA was transferred from 1% agarose gels onto nitrocellulose as described by Southern (29) and hybridized with nick-translated DNA probes as described by Wahl et ale (30). RNA-DNA hybridization and nuclease Sl digestion These were done essentially as described by Berk and Sharp (31) with the following modifications: a) Neither the DNA nor the RNA was radioactive, b) Hybridization was at 60°C for 3 hr, c) Approximately 200 units of nuclease Sl (Sigma) were used for a reaction of 100 ug RNA and 2 ug DNA in a 1 ml volume, and d) Samples were divided in two and run in 1.4% agarose gels in 50 mM Tris pH 8.3, 40 mM NaOCH3, 2 mM EDTA (neutral) or alkaline 1.4% agarose gels (32). Results Synthesis of gpgroES is heat-induced Both Neidhardt and VanBogelen (13) and Yamamori and Yura (14) described a protein of approximately 15,000-Mr whose rate of synthesis was increased by heat-shock. I have independently made similar observations and, furthermore, have shown this protein to be gpgroES as follows. [35SJ-methionine-labeled proteins from normal and heat-shocked bacteria were separated by sodium dodecyl sulfate {NaDodS04)-polyacrylamide gel electrophoresis and the position of the 74 small heat-shock polypeptide was compared with that of gpgroES (Fig. 1). Marker protein was obtained by labeling bacteria containing a multi-copy plasmid including both groE genes (pLS1; see below). The low molecular weight heat-shock protein co-migrates with the gpgroES overproduced in the plasmid-containing strain (Fig. 1). Mixing experiments using unlabeled extracts from the groE-overproducing strain and radioactive extracts of heat-shocked bacteria showed that the two proteins are also found at identical positions after two­dimensional gel electrophoresis {data not shown}. In addition, this protein was found to have identical electrophoretic mobility to that synthesized after infection of UV-irradiated bacteria with a A_groE+ transducing phage {data not shown}. Cloning of the groE genes into plasmid vectors The original cloning of the groE genes into A vectors (6,7,10) involved the selection of groE+ transducing phages from collections of A vectors ligated with HindIII or EcoRI-digested I. coli DNA. One of these transducing phages {made with EcoRI-digested I. coli DNA} was used as a source of DNA for subcloning experiments (Fig. 2), in an effort to a} localize more precisely the groE genes within the groE+ EcoRI fragment of I. coli DNA, b) generate probes for determining the groE gene transcription pattern and c} construct substrates for DNA sequencing of the groE genes and their promoter{s}. The entire 7,900 base pair (bp) groE+ EcoRI fragment of I. coli DNA was first excised from the AgroE+ transducing phage 378 (2) 75 Figure 1. Protein synthesis patterns of normal and heat-shocked bacteria. B178 bacteria were labeled with 10 uCi/ml of [35SJ methionine for 10 min at 30°C (lanes 1,2 and 4) or from 5 to 15 min after a shift from 30°C to 42°C (lanes 3 and 5). The bacteria in lane 1 were transformed with pLSl. The arrows mark the positions of gpgroEL (65,000 Mr) and gpgroES (15,000 Mr ). The percent acrylamide was 12.5. 76 1 2 3 4 5 Figure 2. Construction of plasmids for groE analysis. The!.. col i EcoRI fragment of the A groE+ phage 378 (27) was inserted into the EcoRI site of pBR322 (33) to make pLSl. pLSl was digested with BamHI and EcoRI and religated to make pLl (with 3200 bp of !.. col i DNA) and pSI (with 4700 bp of !.. col i DNA). pSI was partially digested with Sau3A (see Methods and Materials) and ligated with BamHI-digested pJB8 (34). A plasmid with a 2200 bp insert, pS2, was digested with EcoRI and ligated with EcoRI-digested pMOB45 (35). A plasmid with the pS2!.. coli EcoRI fragment inserted into the pMOB45 EcoRI site (pS3) was digested completely with EcoRI and partially with Sau3A and ligated with BamHI-digested pJB8. The resultant plasmid, pS4, has 900 bp of !.. coli DNA inserted into its BamHI site. R- EcoRI; K- JSE.!!.I; B- BamHI. The JSE.!!.I fragments in the .groE+ insert into 378 are approximately 600,3200, [180,95,410 (the order of these fragments is unknown)], and 3760 bp, from left to right. 78 R K KKKK B R + l I + I + I I ).. ---I I , I I I I I I I I I I I I I I i I I I I I I i I I , ,.-- ~ pBR322 R RBR 8amHl + '\V t)+ G EeoRI ligase.. \ pLI 8 pJ88 pLSI 8amHI pS2 pM0845 R R8R '\l/ (:J pJ88 pS3 Bam HI EeoRI Sau3A ligase R 1 c:)/R pS4 by EcoRI digestion and then ligated with EcoRI-linearized pBR322 plasmid (33). A hybrid plasmid generated by these steps (named pLSl) transformed all groE- tested to tetracycline resistance (TcR), ampicillin resistance (ApR), and GroE+. Transformants carrying this plasmid were shown to synthesize high levels of both ~ proteins (Fig. 3, lane 3). The pLSl plasmid was digested with the nucleases BamHI and EcoRI and religated. Two plasmids were generated, one with a 4700 bp and the other with a 3200 bp fragment of I. coli DNA ligated into the hybrid BamHI-EcoRI site of pBR322. The plasmid with the 4700 bp fragment of I. coli DNA was shown to complement groES- but not groEL­bacteria and was named pSl. Bacteria transformed with pSl synthesized increased levels of gpgroES but only normal levels of gpgroEL (Fig. 3, lane 4). The plasmid with the 3200 bp fragment of I. coli DNA was unable to complement any groE- mutants and called pLl. The presence of this plasmid in a bacterial strain does not lead to a detectable increase in the levels of either groE gene product (Fig. 3, lane 2). This plasmid probably contains part of the groEL gene or DNA sequences required for its expression, since the parent plasmid possessed active groEL and groES genes. The pSl plasmid, which probably contains part of the groEL gene, was used as a source of DNA for further subcloning. In order to isolate the smallest possible piece of DNA which includes all the sequences necessary for groES gene expression, pSl was partially digested with the restriction enzyme Sau3A (see Methods and 80 Figure 3. Protein synthesis patterns of bacteria transformed with various plasmids. Bacteria were labeled for 10 min at 37°C with 10 uCi/ml of [35SJ methionine. The bacteria were transformed with no plasmid (1), pLl (2), pLSl (3), or pSI (4). 81 1 2 3 4 1......--,.. Materials) and the resulting fragments were li9ated with BamHI­linearized and bacterial alkaline phosphatase treated pJBB vector (34). The ligation mixture was used to transform KB03, all of the ApR transformants were pooled, and plasmid DNA was isolated from the pool. This DNA was used to subsequently transform groES30 (which transforms about 50-fold less efficiently than KB03) and ApR transformants were screened for their ability to plate A. Among the Gro+ transformants (which simultaneously gained the ability to grow at 42°C) was one which contained a plasmid called pS2, which has - 2200 bp of I. coli DNA ligated into its BamHI site. A problem with the above screen was that both pSI and pJBB confer ApR upon bacteria carrying them, so there was a high background of pSI-derived plasmids that had ligated back to themselves after partial digestion. To circumvent this problem and enrich for recombinant plasmids, pS2 was digested with EcoRI, which excises the insert, and the digestion mix was ligated with EcoRI-digested pMOB45 (35). A hybrid plasmid containing the -2200 bp EcoRI fragment of pS2 was isolated and called pS3. groEs- bacteria carrying this plasmid are TcR, Gro+ and, since the pMOB45 vector has a temperature­sensitive copy control mechanism, such strains synthesize very high levels of gpgroES. The pS3 plasmid was used as a source of DNA for further subcloning into pJBB, since it would not contribute any ApR background. The plasmid was digested to completion with EcoRI (which excises the inserted bacterial DNA) and then partially digested with 83 Sau3A. The mixture was ligated with BamHI-digested pJB8 and the ligation mixture was used to transform K803. All the transformants were pooled and the mixture was used to transform groES30. ApR and Gro+ transformants were screened for the presence of plasmids with small pieces of I. coli DNA inserted into their BamHI sites. A plasmid with an ~900 bp insert was isolated and called pS4. Simultaneous digestion of pS4 with EcoRI (which liberates the insert) and ~I (for which there are no sites in the vector sequences) leads to the formation of DNA fragments ~ 720 and - 180 bp in length, in addition to the linearized vector. These results suggest that the pS4 insert contains one ~I site. Since the - 720 bp EcoRI-~I fragment is larger than any of the ~I fragments in pSI except the - 3200 bp fragment, it must be derived from that piece of DNA. The ~I site in this plasmid, therefore, must be the second one in the groE+ EcoRI fragment of DNA, localizing the groES gene to that region. Transcription of the groE genes The close proximity of the two groE genes and their coordinate induction suggested that they might be co-transcribed. To test this I looked for the presence of a heat-inducible RNA that would hybridize to plasmids into which the ~ genes had been inserted. RNA was prepared from bacteria before and after a 5 min shift to 42°C from 30°C. The RNA was separated by electrophoresis in methyl mercury hydroxide-agarose gels, transfered to diazobenzyloxymethyl paper and hybridized with various probes. The pS2, pLI, and pLSl plasmids all hybridized to an RNA molecule - 2200 nucleotides in 84 length (Fig. 4), although hybridization with pL1 was much less than with the other two probes (see below). As shown in Fig. 4, the levels of this RNA species are increased significantly by 5 min of growth at 42°C. These results suggest that the groE genes are transcribed from a single promoter, whose activity level is changed by heat shock. Mapping the position of the groE transcript The position of the groE transcript on the pL51 plasmid was determined by 51 nuclease mapping. The pL51 plasmid was digested with various restriction enzymes, denatured, hybridized with.I. coli RNA isolated from heat-shocked bacteria, and digested with 51 nuclease. 'The surviving hybrids, which include plasmid DNA hybridized with messenger RNA and some duplex plasmid DNA, were precipitated and separated by electrophoresis in neutral and alkaline gels of 1.4% agarose. Digestion at the single BamHI site in the ~ DNA of pL51 seems to eliminate groEL .gene expression so BamHI­digested pL51 was used as the source of DNA for 51 nuclease mappi ng. Th is experi ment yi e 1 ded fragments of - 2000 and - 2200 bp on the neutral gel and a -2000 nucleotide fragment on the alkaline gel (Fig. 5). These results suggest that transcription begins - 2000 bp from the BamHI site and terminates - 200 bp past it. This model predicts the existence of a - 200 nucleotide fragment on the alkaline gel, but this was not found (see Discussion). '!s'p!!'I-digested pLS1 was used in S1 nuclease mapping experiments to determine on which side of the BamHI site the majority of the groE 85 Figure 4. Hybridization of various plasmids with E. coli RNA. RNA was extracted from bacteria grown at 30°C (lanes 1,4 and 7) or from bacteria that had been heat-shocked for 5 min at 42°C (lanes 2,5 and 8). The RNA was separated on 1.4% agarose-methyl mercury hydroxide gels, transferred to diazobenzyloxymethyl paper and hybridized with pL1 (lanes 1-3), pS2 (lanes 4-6), or pLS1 (lanes 7- 9). Lanes 3,6 and 9 contains BamHI-EcoRI digested pLS1 for size standards. Lanes 1-3 were exposed twice as long so that the hybridization to the transcript would be visible. -86 ABC 123 123 123 Figure 5. Sl Mapping Experiments pLS1 DNA was digested with BamHI (A and B) or ~I (C), hybridized with RNA extracted from heat-shocked bacteria, and digested with Sl nuclease. The surviving duplex nucleic acid fragments were separated by electrophoresis through alkaline (A and C) or neutral (B) gels of 1.4% agarose. The nucleic acids were transferred to nitrocellulose filters and hybridized with nick­translated pLS1 DNA. The autoradiograms shown are the following: (A) alkaline gal and (B) neutral gel of BamHI-EcoRI-digested pLS1 DNA (lane 1) or Sl mapping experiment with BamHI-digested pL51 DNA (lane 2) and (C) alkaline gel of ~I-digested pLS1 DNA (lane 1), 51 mapping experiment with ~I-digested pL51 DNA (lane 2), and linear size standards (lane 3). 88 A B c 1 2 1 2 1 2 3 90 transcript lies. There are 5 ~ sites in pSI, four of them within 2000 bp of the BamHI site, while pLI has no ~I sites. Digestion of pLSI with ~I, hybridization with I. coli RNA. and SI nuclease digestion should, therefore, yield (after alkaline agarose electrophoresis) either 5 fragments of - 410,95,180,750 and 800 nucleotides or a single fragment of - 2200 nucleotides, corresponding to the entire transcript. Contrary to these predictions, an ~ 800 nucleotide fragment of DNA is found on alkaline agarose gels after hybridization and Sl nuclease digestion (Fig. 5). This size corresponds to that of the predicted fragment extending from the ~I site in pSI closest to the BamHI site to a position - 200 bp past that site, suggesting that the groE promoter lies in pSI. The rest of the predicted fragments were not found (see Discussion), although it is likely that the - 750 and - 800 nucleotide fragments co-migrate under the electrophoresis conditions employed. It appears that only ~ 20~ nucleotides of pLI are transcribed into groE RNA. This may be the reason that hybridization of nick­translated pLI with total I. coli RNA immobilized on diazobenzyloxymethyl paper was much weaker than that found with either pS2 or pLSI (Fig. 5). Discussion The results described in this chapter suggest that the groEL and groES genes are co-transcribed from a promoter located about 2000 bp to the pSI side of the BamHI site in the groE genes and that transcription of the operon is induced by heat-shock. The direction of this transcription is most likely promoter -groES-groEL. The following results support this model: a) the pS1 plasmid contains a functional groES gene and the synthesis of gpgroES from the plasmid gene is heat-induced (K. Tilly, unpublished results), suggesting that transcription is from the normal groE promoter, b) DNA sequences necessary for groEL gene expression are located on either side of the BamHl site in the groE+ EcoRl fragment of I. coli DNA, c) polar mutations (caused by insertion of either Tn5 (36) or the Mud(lac) phage (37) can inactivate the expression of both genes simultaneously although the insertion mutations were not structurally characterized (K. Tilly, unpublished results), and d) the rates of groES and groEL protein synthesis in response to many changes in the physiological state of the bacteria are exactly parallel to each other (F. Neidhardt and R. VanBogelen, personal communication). The results of the Sl mapping experiments are consistent with the co-transcription model but several predicted fragments were not found. When BamHl-digested pLS1 was used as the DNA source, the predicted 200 nucleotide fragment of protected DNA was not found after alkaline agarose gel electrophoresis. Similarly, the predicted 750,410,95 and 180 nucleotide fragments were not found when ~l-digested pLS1 was used as the DNA source. It is likely that the 750 and 800 nucleotide fragments migrate together under alkaline agarose electrophoresis since the difference in their sizes is within the error of the restriction digest mapping. The absence of the small fragments could stem from either or both of the following reasons. First, the small fragments 91 may have been lost in the transfer of DNA to nitrocellulose filters (especially since the DNA was depurinated before transfer). Second, hybridization to these fragments may have been poor, either simply as a consequence of their small sizes or because they form diffuse bands under the electrophoresis conditions used. Regardless of the polarity of groE gene transcription, it is clear that the genes are co-transcribed and co-regulated. Previous studies have shown that both ~ gene products act in A head assembly and T5 tail assembly. Both proteins appear to act at an early step of head and tail assembly, as judged by the phenotype of mutations in either gene (1,2,4,11). Furthermore, it has been shown that the two gene products interact functionally, since suppressors of mutations in the groES g~ne can be found in the groEL gene (23). Because the ~ gene products appear to act together in these processes, it might be advantageous to the cell to co-regulate their synthesis, in order to maintain the appropriate ratios of the two proteins. The groE gene products are also important for bacterial metabolism, since many groE- mutants are temperature-sensitive for growth and since the synthesis of the groE gene products is induced by various stimuli. In most cases, the control of the two groE genes is clearly a function of the positive regulatory gene ~. Induction of gpgroEL synthesis by amino acid starvation (22,38) may be an exception to this rule, since synthesis of the rest of the heat shock polypeptides has not yet been shown to increase, but C. Woolford (personal communication) has found that gpgroES synthesis 92 may also be'induced in those conditions. If induction is independent of htpR, there may be an additional control region or a secondary promoter (for at least the groEL gene) at which control by amino acid starvation is mediated. The inclusion of the groE genes in the group of genes regulated by the htpR gene product suggests that their products may play important roles in bacterial survival of the face of stress. How heat-shock proteins work together to promote survival of stresses is not und