ABSTRACT. RegA protein of T4 and related bacterio- phages is a highly conserved RNA-binding protein that re- presses the translation of many phage mRNAs ...
Proc. Nati. Acad. Sci. USA Vol. 89, pp. 5053-5057, June 1992
Genetics
Regions of bacteriophage T4 and RB69 RegA translational repressor proteins that determine RNA-binding specificity CATHERINE E. JOZWIK AND ERIC S. MILLER* Department of Microbiology, North Carolina State University, Raleigh, NC 27695-7615
Communicated by H. E. Umbarger, February 3, 1992 (received for review November 6, 1991)
ribosome loading (9-12). RegA has not been shown to contain any known RNA-binding motif, and it binds with differential affinity to target mRNAs that have no discernible consensus sequence or structure (12-15). To identify amino acids of the RegA protein that are involved in RNA binding, we examined regA from several even-numbered T phages (T-even phages) and RB phages. RB bacteriophages are T-even-like in that they contain hydroxymethylated cytosine, are inactivated by sera against T-even phages, and can complement a number of T-even phage mutations (16). Although regA is not essential for T4 development in the laboratory (9), the RegA amino acid sequences from all of the T-even and RB phages examined previously are identical (17). Preliminary regA sequence and expression data suggested that RB69 encodes a RegA protein with potentially interesting structural and RNA-binding properties. The present study describes the cloning, sequencing, and functional characterization of RB69 regA.t We have also determined the nucleotide changes of several T4 and RB69 regA mutations and have characterized the mutant proteins using an in vivo translational repression assay. Together, phylogenetic and mutational analyses suggest that the aminoproximal portion of the protein governs the recognition of specific mRNAs by RegA.
ABSTRACT RegA protein of T4 and related bacteriophages is a highly conserved RNA-binding protein that represses the translation of many phage mRNAs that encode enzymes involved in DNA metabolism. RB69, a T4-related bacteriophage, has a unique regA gene, which we have cloned, sequenced, and expressed. The predicted amino acid sequence of RB69 RegA is 78% identical to that of T4 RegA. Plasmidencoded RB69 RegA expressed in vivo represses the translation of T4 early mRNAs, including those of rIIA, rIB, 44, 45, rpbA, and regA. Nucleotide sequences were determined for several T4 and RB69 regA mutations, and their corresponding repressor properties were characterized. AU of the 10 missense mutations affect residues conserved between RB69 and T4 RegA. Two regions of RegA are especially sensitive to mutation: one between Val-15 and Ala-25 and another between Arg-70 and Ser-73. Sequence alignments and mutational data suggest that the region from Val-15 to Ala-25 is similar to helix-turn-helix domains of DNA-binding proteins and confers RNA-binding specificity upon RegA. The RegA691 protein (Ile-24 -- Thr) has an in vivo phenotype that appears to distinguish site-specific and cooperative binding modes of hierarchical RegA-mediated translational repression.
Mechanisms of RNA binding by proteins are not well understood, and few protein motifs involved in RNA binding have been identified. The best characterized motif is the ribonucleoprotein consensus sequence, which consists of 90 moderately conserved amino acids and a highly conserved octapeptide. It is found in a number of ribonucleoproteins and poly(A)-binding proteins (1). Another recently identified RNA-binding motif is the arginine fork, which is composed of a critical arginine residue surrounded by several basic amino acids. Although arginine-rich regions are found in many RNA-binding proteins, the arginine-fork has only been studied in the human immunodeficiency virus (HIV) Tat/Rev transactivator proteins (2). Prokaryotic translational repressors have provided excellent model systems for investigating the roles of RNA structure and sequence in determining RNA-binding specificity. The phage R17 coat protein represses translation of the replicase message by binding to a simple RNA hairpin that sequesters the ribosome binding site (3). A simple RNA hairpin is also recognized for autogenous translational control by phage T4 DNA polymerase (4, 5). Escherichia coli ribosomal protein S4 binds to a double RNA pseudoknot structure to repress a operon mRNA translation (6), and a pseudoknot is required for T4 gene 32 protein to bind its mRNA (7). One of the most intriguing translational operators occurs in the E. coli threonyl-tRNA synthetase mRNA, whose 5' leader region assumes a tRNA-like structure (8). Bacteriophage T4 RegA is a small, slightly basic protein that represses the translation of several T4 early mRNAs by binding to the translation initiation region and preventing
MATERIALS AND METHODS Strains, Plasmids, and Media. E. coli B-strain NapIV (18) and bacteriophage T4D were used throughout. T4 strains of genotypes regAR9, 33amN134, 44amN82, SSamBL292, 62amE1140 (9); regAmd67 (19); and regAmd73 (19); and the phage A cI857-PL (leftward promoter) expression vectors pJM9 (cI857-PL), pJM1403 (c1857-PL:regA), and pTH4553 (cI857-PL:45.2, rpbA, 45, 44) were obtained from J. Karam (20, 21). The constructs pEM104 (cI857-PL:regA104) and pJM1168 (cI857-PL:45.2, rpbA, 45, 44, 62, regA1168) have been described (13, 22). pWK105 was obtained from W. Kelly. T4 strains regA9, regAl), and regA15 were from J. Wiberg (23, 24). Bacteriophage RB69 was obtained from W. Wood. All culture media have been described (13). PCR, Cloning and Sequencing. Bacteriophage genomic DNA was amplified from a single plaque essentially as described (17). Primers used for amplification and nucleotide sequence analysis of RB69 regA were P62 (17) or P62-69 (5'-GAACTGAAGGCCATGGTC-3') with P69 (5'-AGACGAGTTATTGTCTTTTGTTTG-3'; ref. 5) or P69-2 (5'CGTTTCTTTAAAGTGGAC-3'). For amplification and sequence analysis of T4 regA, primers P62 and P5 were used (17). All regA alleles were PCR-amplified, and the fragments were treated with Klenow polymerase (25) and ligated into pBluescript KS(+). A BamHI/Pvu II fragment was then transferred to the BamHI/Pvu II sites of pJM9. The resulting Abbreviations: T-even phage(s), even-numbered T phage(s); PL, phage A leftward promoter. *To whom reprint requests should be addressed. tThe sequence reported here has been deposited in the GenBank data base (accession no. M86231).
The publication costs of this article were defrayed in part by page charge payment. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. §1734 solely to indicate this fact. 5053
5054
Proc. Natl. Acad. Sci. USA 89 (1992)
Genetics: Jozwik and Miller
cI857-PL constructs (heat inducible) were designated pEC69, which contains the RB69 wild-type regA gene, and pEC691, pEC693, pEC695, and pEC697, which contain the corresponding mutant RB69 regA alleles. T4 genes 45.2, rpbA, 45, and 44 were transferred from pTH4553 to pWK105, which contains the R6K origin of replication and the Tn9O3 kanamycin-resistance gene (26). The resulting plasmid, pCJ4000, has these T4 genes under phage A cI857-PL transcriptional control. Cloning procedures were as described (27). Nucleotide sequences of T4 and RB69 regA mutations were determined from asymmetrically amplified phage DNA or from single-stranded DNA of pBluescript (Stratagene) regA clones (17, 28). Sequence analysis was done with Sequenase Ver. 2.0 by following the manufacturer's instructions (United States Biochemical). Computer programs were from the Genetics Computer Group (29). In Vivo Labeling and Translational Repression Assays. Plasmid-bearing E. coli B cells were grown (300C) to 3 x 108 cells per ml in M9 minimal-glucose medium that contained 0.02% Casamino acids and ampicillin at 50 pug/ml. Proteins were labeled for 4 min (at 30°C or 42°C) with 10 ,Ci (370 kBq) of cysteine/methionine Tran35S-label (ICN Biomedicals) as described (13). For repression assays, plasmid-containing cells were grown to 3 x 108 cells/ml and shifted to 42°C for 15 min; then strain T4 regAR9, 33amN134, 44amN82, SSamBL292, 62amE1140 was added at a multiplicity of infection of seven. Five minutes after infection, 10 ,Ci of Tran35S-label were added, and incorporation of isotope was stopped 10 min later (13). For cells containing compatible plasmids encoding RegA and T4 target mRNAs, cultures were grown, induced, and labeled as above except that phage were not added. Proteins were analyzed by SDS/polyacrylamide gel electrophoresis followed by autoradiography or immunoblot (Western) analysis (13). Autoradiographs were analyzed on a LKB UltroScan XL laser densitometer (Pharmacia LKB Biotechnology). Isolation of RB69 and T4 regA Mutations. Several T4 regA mutations were previously isolated and characterized in the phage as described (19, 23, 24, 30). T4 regA mutations on pEM31, pEM32, pEM33, pEM34 and pEM35 were isolated as spontaneous mutations in the regA plasmid pEM302 (cI857-PL:regA; ref. 13). To enrich for regA mutations, an overnight culture of NapIV/pEM302 was diluted, spread onto enriched Hershey agar containing tetracycline (10 ,ug/ ml) and incubated at 37°C for 2-3 days. Twenty larger colonies were selected as possible regA mutants. RB69 regA mutations were isolated as Taq DNA polymerase errors under standard symmetric PCR amplification conditions. Mutant alleles were cloned into pJM9 as described for wild-type RB69 regA, and their sequences were determined.
RESULTS Cloning and Nucleotide Sequence of RB69 regA. Fig. 1 presents the RB69 regA nucleotide sequence that was determined from cloned and asymmetric PCR products. RB69 is the only phage examined to date in which regA is significantly different from that in T4. It is 73% identical to T4 regA, and there are no insertions or deletions relative to the 366nucleotide T4 regA coding region (22). Immediately outside the coding region, RB69 has an additional A at the regA translation initiation and autogenous control site. Insertions and deletions 3' of regA conserve the stem-loop structure, maintaining the extreme stability observed for this type of hairpin (32). RB69 RegA Amino Acid Sequence and Expression. The predicted sequence of the 122-amino acid RB69 RegA protein is 85% similar (78% identical) to T4 RegA, with the greatest degree of similarity (93%) occurring in the amino-terminal half (Fig. 1). Residues in the carboxyl-terminal half display only 77% similarity to the T4 protein.
62 *** *** red GAACGTAAAAGAACAGAAACAACTCAAAAAACTAGCATTGGAATGGTAA.AATGATTGAA T4 A T RB69 A T G A T T
A
T4 ATTACTCTTAAAAAAC RB69 AAT G G C T4
RB69
C
G
C
59 60
GA 119 C 120
C
ATTGCTAATAATAAAGATAAAGTTCTATATCAGTCCTGTCATATTCTTCAGAAAAAAGGA 179 c
C G
T
C
C
A
GC
T180
T4 CTATACTATATCGTTCATTTTI AAMGAAATGC7TTCGTATGGATGGTCGTCAAGTTGAAATG 239 C C RB69 AA T C C A T T 240 G G T4 ACAGAAGAAGATGAAGTTCGTCCGTGATTCGA'TTGCATGGCTATTAGAAGATTGGGGACTA 299 T G 300 TCA T GC T RB69 GAT GT CT TCAG A C
ATTGAAATCGTTCCTGGTCAAAGAACTTTTATGAAAGATITAACTAATAACTTCCGAGTT 359 T4 360 T TT T AAGACTC GCTC TGAAGA C TTT GCC G RB69 ATTAACGTTCCTAAATATACGATTGGTAATTAA 419 T4 420 A CTT GAAAG A C C G G RB69 T4 . GCAA.GGGGTTCGeCCCTTATTTGGAGTAT 450 .. 451 A T RB69 A v
B T4
NIEITLKKPEDFLKVKETLTRGIANNKDKVLYQSCHILOKKGLYYIVHFIKELRYDGRQV
61 61
RB69 ----K--N-
Y
Ivt
h1i AQL C m1 I81 C* L ENTnEDEVRRDSIAWLLEDWGLIEIVPGQRTFNKDLTNNFRVISF QKiUKLVPKYTIGN 122 122 -DD-T*EDSA-EDLFG-RB69 DIDGT4
V
FIG. 1. Aligned RB69 and T4 regA nucleotide (A) and amino acid (B) sequences. The initiation and termination codons (* above the residue), CUUCGG hairpin (-*), and RegB (31) RNA cleavage site (#) are indicated. Amino acids of identity (-), conservative change (:), and limited conservation ( ) used in the similarity comparison are indicated. Substitutions and alterations in mutant proteins are shown above (T4) and below (RB69) the affected residue. Fonts used for mutant proteins are: lightface, mutations with wild-type activity; underlined, mutants with altered activity; boldface, mutants with no activity. Double mutants are *, regA91; t, regA695; and QL, regA671.
RB69 regA was placed under the transcriptional control of PL in the pEC69 construct. Induction of RB69 regA transcription followed by Western blot analysis demonstrated the synthesis of a protein that reacted with rabbit anti-T4 RegA polyclonal antiserum (Fig. 2). Antibody binding to RB69 RegA is reduced relative to the reactivity observed with T4 RegA. The slightly faster migration of RB69 RegA, relative to T4 RegA, in SDS/PAGE is consistent with the lower calculated mass (14.4 vs. 14.6 kDa) and pI (7.5 vs. 9.6). Fig. 2 also shows that expression of RB69 regA is autogenously controlled since the regA null mutations contained on pEC695 and pEC697 (Table 1) caused a nearly 10-fold increase in the production of the protein. Additionally, when cells were induced and incubated with [35S]cysteine/methionine, the wild-type RB69 regA expression vectors consistently produced 50-67% less RegA protein than identically configured T4 regA clones (data not shown). Repression of T4 mRNAs by RB69 RegA. Repression of phage-encoded messages was assayed by infecting cells that
-.
FIG. 2. Synthesis of RegA from RB69 regA plasmids. Cells containing cI857-PL:regA constructs with the wild-type (T4+, pJM1403; RB69+, pEC69) and RB69 mutant (regA691-regA697) alleles were subjected to Western blot analysis with anti-T4 RegA polyclonal antiserum as described (13). Note that RegA691 retains autogenous control.
Genetics: Jozwik and Miller contained plasmid-encoded RB69 RegA with T4 regA phage (17). Alternatively, a two-plasmid system was used in which one plasmid afforded temperature-inducible regA expression and the other plasmid, pCJ4000, contained the T4 RegAsensitive genes 44, 45, and rpbA. Results of SDS/PAGE analysis of pulse-labeled proteins showed that RB69 RegA represses the same mRNAs as T4 RegA, including those encoded by rIIA, HIB, 44, 45 and rpbA (Figs. 3 and 4). Synthesis of wild-type RB69 RegA consistently caused 60% less incorporation of radiolabeled amino acid into protein during T4 infection (Fig. 3), as determined by scanning laser densitometry. Neither wild-type T4 nor mutant RB69 RegA inhibited overall phage protein synthesis to this extent. T4 and RB69 regA Mutations. To identify regions required for RNA-binding specificity, regA was analyzed as follows. (i) We determined the nucleotide sequence of the previously isolated T4 mutant alleles regAR9, regA9, regAlI, regA5, regA104, regAmd67, and regAmd73 (13, 19, 22-24, 30). (ii) An E. coli culture containing the wild-type phage A cI857PL:regA configuration (pEM302) was enriched for T4 regA Table 1. Nucleotide changes and translational repression properties of regA mutants Allele Mutation* Alteration Activityt T4 (phage) regAJ1f G-49 A Glu-17 - Lys Altered regA73 A-52 G Thr-18 - Ala Altered (rIIB, regA) regA9O C-74 A Ala-25 - Asp Altered regA67 C-218 T Ser-73 - Leu Altered (45, regA) A A-121 regAR9 Gly-43 FS§ Null A G-242 regA15 Gly-82 FS§ Null T4 (plasmid) Val-15 Leu Wild type G-43 T regA9OR C-74 - T Ala-25 Val Wild type regA34 C-74 A Ala-25 Asp Null regA91 C-208 T Arg-70 Cys regA671 A-215 G Asp-72 Gly Altered (rIIA, 45, regA) C-218 T Ser-73 Leu regA104 G-336 - C Trp-112 - Cys Null regA33 G-7 - T Glu-3 -- Stop Null regA31 G-336 ISliI Trp-112 - ISi1 Null regA32 A-350 ISI1 Lys-117 - IS1I Altered (regA) G-193 ISSl Thr-64 IS1 No data regA35 RB69 (plasmid) T-71 C Ile-24 -o Thr Altered (rIIA, regA691 rIB, 44, 45) C-74 C regA69S Ala-25 - Val Null C-109 - T His-37 - Tyr A-215 C regA697 Asp-72 Gly Null regA693 C-289 G Leu-97 - Val Wild type All nucleotide sequences were determined in this work with the exception of that of regA9OR which was previously characterized
Proc. Natl. Acad. Sci. USA 89 (1992)
5055
300C 420C 69+ 697 T4 T4+ 69+ 691 693 695 697
M T4+
- ;-- _-*- -q_
t
v-
_.~~-
-
-..
-~-
-
=4
-r
V~I
RpbA ~~~~~~~ _ 45
62 RpbA
-rlIA
E
.1B
:. : _4
_
_
f;;;
$
.
=
:*
X
FIG. 3. Translational repression of T4 early protein synthesis by plasmid-encoded RB69 RegA. Cells containing induced (42°C) levels of RB69 RegA were infected with T4 regAR9 phage, and proteins were analyzed as described. *, Repressed. Wild-type regA alleles are on expression plasmids pJM1403 (lanes T4+) and pEC69 (lanes RB69+); lane T4- refers to T4 regA104 on pEM104, and all other lanes refer to constructs pEC691-pEC697, which contain the respective RB69 regA mutations. T4 marker proteins (lane M) are from pJM1168.
mutations (ref. 13; see Materials and Methods). Nucleotide sequence analysis showed that of 20 isolates examined, 5 had mutations in regA (regA31, regA32, regA33, regA34, and regA35); all others had the wild-type regA sequence. (iii) PCR amplification and subsequent cloning of regA from phage or plasmid DNA produced four missense alleles from RB69 (regA691, regA693, regA695, and regA697) and two from T4 (regA91 and regA671). The sequences and repression activity of the 19 RegA mutant proteins studied are summarized in Table 1. The mutant RegA proteins exhibited a range of translational 420C
30°c regA: pCJ4000:
-
69+ -
+
-
+
T4- T4t 69+ 691 693 695 + + + + + +
697 +
_~~~~~~
M
44
(33).
*Mutations are numbered with the A of the regA initiator AUG as nucleotide 1. tWild-type activity, repression of rIIA, rIIB, 44, 45, rpbA, and regA; altered activity, repression of some targets but not others (parentheses indicate those genes that are not repressed); null activity, complete lack of repression. tRegA9 and RegAll partially repress most targets after extended synthesis (23). §The regAR9 frameshift (FS) arose by deletion of an A from the sequence GAAAAAAG beginning at nucleotide 120, and the regA15 frameshift arose by deletion of a G from the sequence TGGGGA beginning at nucleotide 241. ISI1 (intervening sequence 1) insertions are at the indicated nucleotide.
45
Rpb
FIG. 4. Repression by RB69 RegA of plasmid-encoded T4 mRNAs of 44, 45, and rpbA. Target mRNAs are expressed from pCJ4000 in E. coli B containing a compatible plasmid with the indicated regA allele (see the Fig. 3 legend). Proteins were analyzed as described in Materials and Methods. *, Repressed.
5056
Genetics: Jozwik and Miller Helix
-
Turn
Proc. Natl. Acad. Sci. USA 89 (1992) Helix
-
DNA-binding ACI ACro 434Cro
434cI GalR NifA NtrC
OmpR DctD
E T T A K A
S V A D Q Q KT A K E L A T Q E L A Q Q I D V A R K A A R Q K Q E A A R G E E V D R V K E T L Q
K
D K K L L L I A
M
L A V A L L V L
G G G G G G G G G
N
V V T V N
W L I
G Y K T S T G E P
Q S G V G A L Q S A I N KA Q Q S I Q L I Q Q S I E Q L VATVSRV P R Q V A Y R R N T L T R K I G A D D Y I R K T F Y D K
&
N R L Y
F I E E I I L P L
N H A N N Q K K
G I A G G V G K N S IN E L P F Q RH
N R T T P D G N G
V L Y Q K F HH V G F I Q L C Y
S C i S E _ AN H V D L
DISCUSSION
RNA-bfdUmg
RegA MtrB
BgIG
WpV
V V G T
K E E R
Z I L T R M GI D G V N V E A L T*R S G V S R
72% similar (55% identical) in the 11-residue helix-turn region. Overall, RegA is highly similar to all of the DNAbinding proteins in the turn region, including the invariant glycine, and shows significant similarity in the first helix. Applying Chou and Fasman's rules for predicting protein secondary structure (38), the amino-terminal portion of Reg A is likely to form an amphipathic a-helix prior to the turn.
I GL T L G V Q Q G K P
N G P S
K T K V
DK D T D E N E
FIG. 5. Alignment of helix-turn-helix regions of prokaryotic DNA-binding proteins to RegA and other RNA-binding proteins. Underlined residues in RegA and MtrB mark mutated sites (this work and ref. 34). Sequences of the phage transcriptional repressor proteins were aligned as described (35). Other sequences were extracted from the SwissProt (Sw) or GenBank (Gb) data bases by the following accession numbers: Bacillus subtilis MtrB, Sw:P19466; E. coli BgiG, Sw:P11989; GalR, Sw:P03024; NtrC, Sw:P06713; OmpR, Sw:P03025; Klebsiella pneumoniae NifA, Sw:P03027; Rhizobium meliloti DctD, Sw:P13632; fl gpV, Gb:J02450. For BgiG, the sequence IIDK, ca. one a-helical turn, was omitted at the asterisk to facilitate presentation.
repression phenotypes, including no appreciable activity (e.g., RegA15, RegA104, and RegA695), altered activity (e.g., RegA73 and RegA691), and wild-type activity (e.g., RegA34 and RegA693). Repression properties of several regA mutants isolated in T4 have been reported, including some altered mRNA target specificity mutants (19, 23). We present the in vivo activities of cloned RB69 regA mutants. From Fig. 2 it is apparent that RegA691 and RegA693 retain significant autogenous control of translation from plasmid-encoded regA mRNA. In contrast, RegA695 and RegA697 have lost this property. The null phenotype of these two proteins is apparent from phage infections or two-plasmid assays, in which they fail to repress all RegA-sensitive mRNAs (Figs. 3 and 4). RegA693, on the other hand, has apparent wild-type activity on all target mRNAs. regA691 is an interesting mutation that results in a RegA protein with repressor activity on regA message (and on another mRNA encoding an unidentified protein; Fig. 3) but lacks repressor function on the other mRNAs. Alterations in each of the T4 and RB69 mutant RegA proteins are shown in Fig. 1. Nucleotides specifying those regions of the protein between Val-15 and Ala-25, between Arg-70 and Ser-73, and near Trp-112 are particularly sensitive to mutation. Most notable are the nucleotides corresponding to the Val-15-toAla-25 region, which have 7 of the 10 single missense mutations, 4 of which affect target specificity (regAlI, regA73, regA9, and regA691; Table 1 and Figs. 2-4). This interval between Val-15 and Ala-25 is contained within the longest conserved stretch of amino acids in T4 and RB69 RegA. Amino acid sequence conservation in T4 and RB69 RegA, and the distribution of mutations that alter the pattern of repression, implicate the region between Val-15 and Ala-25 in conferring RNA-binding specificity to RegA. Although not altered as often, the conserved residues from Arg-70 to Ser-73 (Table 1) may also constitute part of the RegA RNA-recognition site. The apparent site-specificity region between Val-15 and Ala-25 was used in FASTA (29) alignment searches of the Swiss-Prot and GenBank data bases (Fig. 5). Surprisingly, similarity to the helix-turn-helix motifs of several prokaryotic DNA-binding proteins was revealed (Fig. 5), including phage repressor proteins (35), DctD, NifA, NtrC, and OmpR (36). The similarity of RegA and the latter class of proteins is most striking for R. meliloti DctD (37); the two proteins are
The predicted 122-amino acid sequence of RB69 RegA is 85% similar to that of T4 RegA. Sequence alignment in Fig. 1 shows that the amino-terminal portion of the proteins is highly conserved, containing only 4 of the 26 amino acid differences between the two proteins. This conservation, compared with the numerous substitutions in the carboxylterminal portion, suggests that the amino terminus of RegA is important for RNA recognition. Plasmids with RB69 regA under control of the phage A c1857-PL system afford inducible synthesis of RB69 RegA protein. A comparison of the synthesis of RegA from wildtype and mutant alleles (Fig. 2) demonstrates that RB69 regA is translationally -autoregulated and is regulated more efficiently than is T4 regA. The relative contributions of the nucleotide changes near the translational operator or the amino acid differences in the repressor protein to this more stringent autogenous control have yet to be determined. RB69 RegA protein repressed all of the T4 RegA-sensitive mRNAs. In the phage repression assays, T4 regA strain infections of cells containing wild-type RB69 RegA displayed a typical repression pattern for specific T4 mRNAs. An overall diminution of phage protein synthesis was also noted (Fig. 3). Both activities were lost by proteins encoded by the nonautoregulated alleles regA695 and regA697. The in vivo repression properties of RB69 RegA clearly show that it specifically and generally inhibits protein synthesis more extensively than T4 RegA. Mutationally sensitive regions of the RegA protein were identified by determining the nucleotide sequence of several existing and newly isolated regA mutations (Table 1). Our results show that all of the missense mutations affect residues conserved in the T4 and RB69 RegA proteins (Fig. 1B). Two regions of the RegA proteins are especially sensitive to mutation, the most notable being between Val-15 and Ala-25, which showed similarity to the helix-turn-helix domains of several DNA-binding proteins. Other RNA-binding proteins also show sequence similarity to the region of RegA resembling the helix-turn-helix motif (Fig. 5), including the B. subtilis MtrB protein, which appears to regulate attenuation in the trp operon by binding to a hairpin in the RNA leader (34). Mutations affecting translational repression by RegA and attenuation control by MtrB affect residues in the aligned region. The turns of both proteins appear to be most sensitive; MtrB-negative complementing mutants are located here (34), as are RegA site specificity mutants, such as regA9 and regA691 (Table 1 and Fig. 5). Phage fl gpV translational repressor protein (39) and E. coli BglG antitermination protein (40) also have a region of similarity to the helix-turnhelix motif. A recent NMR study of the RegA binding site on T4 44-encoded mRNA revealed that this RNA is single-stranded and that the phosphodiester torsion angles and the ribose 2'-endo puckering resemble that of a B-form DNA double helix (14). Thus, RegA may use features of the wellcharacterized helix-turn-helix motif to recognize singlestranded RNA, which may be a mechanism used by other
RNA-binding regulatory proteins. When RegA is added to in vitro transcription/translation reactions expressing RegA-sensitive genes, low concentrations of protein repress the translation of the 44 and rpbA messages, intermediate RegA concentrations repress the
Proc. Natl. Acad. Sci. USA 89 (1992)
Genetics: Jozwik and Miller 44 Q
speificity cooperativity
++++ -
rpbA CXm rlIB m +++
++
regAl CM +
+
FIG. 6. Model for hierarchical RegA-mediated translational repression. Contributions of site-specific and cooperative interactions determine translational repression efficiency. Heavy lines, mRNA translation initiation region; circles, RegA monomers. The representation of bound RegA molecules and the relative contributions of specificity and cooperativity are based on the mRNA protected by RegA against RNase digestion and the repression characteristics of RegA691.
rIIB and 45 transcripts, and high concentrations of RegA repress the translation of the regA message (12). It is not known how this hierarchy of translational repression is established. However, properties of the RB69 RegA691 protein (Ile-24 -* Thr mutation in the proposed specificity domain) may provide insight into RegA-RNA interactions. RegA691 has lost the ability to control most of the T4 target mRNAs but surprisingly, it still represses the regA (Figs. 2-4) and rpbA (Fig. 4; preliminary results) messages. This leads us to propose that a combination of sequence specificity and cooperative protein-protein interactions explain not only the repression of the rIIB message as suggested by Webster and Spicer (15), but the overall hierarchical repression by RegA as well. In our model (Fig. 6), some RegA targets promote strong specific interactions (e.g., 44), others show weak specific interactions but support strong cooperative binding (e.g., regA), and others are repressed by different combinations of sequence specificity and cooperativity (e.g., rIIB and rpbA). Thus, it appears that the mutant RegA691 protein lacks specificity for sequences present in the better targets (i.e., 44) but may retain polynucleotide affinity and the cooperative binding parameter. Repression of the regA and rpbA mRNAs by RegA691 is consistent with this proposal. regA repression is mediated largely through cooperative interactions; thus, binding of RegA691 to this mRNA is not expected to be significantly altered. The rpbA mRNA is normally repressed by the high-affinity binding and extensive cooperative interactions of wild-type RegA. We predict that RegA691 still represses rpbA but less efficiently than the wild-type protein because it lacks the high sequencespecificity parameter. The intermediate strength targets (i.e., r1IB and 45) are not controlled by RegA691 because the cooperativity does not compensate for the lack of specific interactions. Further support is provided by the observations that RegA has an affinity for and binds cooperatively to poly(U) and poly(dT) (15, 41) and that RegA protects varied lengths of the sensitive mRNAs from nuclease digestion (10-12), suggesting that target mRNAs may be bound by one or multiple RegA molecules. RB69 RegA and the mutant proteins described in this study should be useful for further biochemical and structural analysis of RegA and RegA-RNA interactions. We are grateful to J. Karam and J. Wiberg for sending us many of the phage and plasmids used in this work; to J. Gaughran, M. Mammarella, and M. Singer for timely contributions to the sequence and mutant analysis; and to B. Gillette, C. Hardin, H. Kreuzer, T. Petty, K. Tatchell, and S. Udell for thoughtful discussions. This work was supported by Grant GM38659 from the National Institutes of Health. 1. Bandziulis, R. J., Swanson, M. S. & Dreyfuss, G. (1989) Genes Dev. 3, 431-437. 2. Calnan, B. J., Tidor, B., Biancalana, S., Hudson, D. & Frankel, A. (1991) Science 252, 1167-1171. 3. Romaniuk, P., Lowary, P., Wu, H.-N., Stormo, G. &r Uhlenbeck, 0. (1987) Biochemistry 26, 1563-1568. 4. Andrake, M. D. & Karam, J. D. (1991) Genetics 128, 203-213.
5057
5. Tuerk, C., Eddy, S., Parma, D. & Gold, L. (1990) J. Mol. Biol. 213, 749-761. 6. Tang, C. K. & Draper, D. E. (1989) Cell 57, 531-536. 7. McPheeters, D. S., Stormo, G. D. & Gold, L. (1988) J. Mol. Biol. 201, 517-535. 8. Moine, H., Romby, P., Springer, M., Grunberg-Manago, M., Ebel, J. P., Ehresmann, B. & Ehresmann, C. (1990) J. Mol. Biol. 216, 299-310. 9. Karam, J., Gold, L. & Singer, B. S. (1981) Proc. Natl. Acad. Sci. USA 78, 4669-4673. 10. Unnithan, S., Green, L., Morrissey, L., Binkley, J., Singer, B., Karam, J. & Gold, L. (1990) Nucleic Acids Res. 18, 7083-7092. 11. Webster, K. R., Adari, H. Y. & Spicer, E. K. (1989) Nucleic Acids Res. 17, 10047-10068. 12. Winter, R. B., Morrisey, L., Gauss, P., Gold, L., Hsu, T. & Karam, J. (1987) Proc. Natl. Acad. Sci. USA 84, 7822-7826. 13. Miller, E. S., Karam, J., Dawson, M., Trojanowska, M., Gauss, P. & Gold, L. (1987) J. Mol. Biol. 194, 397-410. 14. Szewczak, A. A., Webster, K. R., Spicer, E. K. & Moore, P. B. (1991) J. Biol. Chem. 266, 17832-17837. 15. Webster, K. R. & Spicer, E. K. (1990) J. Biol. Chem. 265, 9007-9014. 16. Russell, R. L. & Huskey, R. J. (1974) Genetics 78, 989-1014. 17. Miller, E. S. & Jozwik, C. E. (1990) J. Bacteriol. 172, 51805186. 18. Nelson, M. A., Ericson, M., Gold, L. & Pulitzer, J. F. (1982) Mol. Gen. Genet. 188, 60-68. 19. Karam, J., Dawson, M., Gerald, W., Trojanowska, M. & Alford, C. (1982) in Interactions of Translational and Transcriptional Controls in the Regulation of Gene Expression, eds. Grunberg-Manago, M. & Safer, B. (Elsevier Biomedical, New York), pp. 84-89. 20. Liang, Y., Wei, R., Hsu, T., Alford, C., Dawson, M. & Karam, J. (1988) Genetics 119, 743-749. 21. Hsu, T., Wei, R., Dawson, M. & Karam, J. D. (1987) J. Virol. 61, 366-374. 22. Trojanowska, M., Miller, E. S., Karam, J., Stormo, G. & Gold, L. (1984) Nucleic Acids Res. 12, 5979-5993. 23. Cardillo, T. S., Landry, E. F. & Wiberg, J. S. (1979) J. Virol. 32, 905-916. 24. Wiberg, J. S., Mendelsohn, S. L., Warner, V., Aldrich, C. & Cardillo, T. S. (1977) J. Virol. 22, 742-749. 25. Scharf, S. J. (1990) in PCR Protocols: A Guide to Methods and Applications, eds. Gelfand, D. H., Innis, M. A., Sninsky, J. J. & White, T. J. (Academic, San Diego), pp. 88-91. 26. Kelly, W. (1991) Ph.D. thesis (Duke University, Durham, NC). 27. Maniatis, T., Fritsch, E. F. & Sambrook, J. (1982) in Molecular Cloning: A Laboratory Manual (Cold Spring Harbor Lab., Cold Spring Harbor, NY). 28. Brow, M. D. (1990) in PCR Protocols: A Guide to Methods and Applications, eds. Gelfand, D. H., Innis, M. A., Sninsky, J. J. & White, T. J. (Academic, San Diego), pp. 189-1%. 29. Devereux, J., Haeberli, P. & Smithies, 0. (1984) Nucleic Acids Res. 12, 387-395. 30. Karam, J. D. & Bowles, M. G. (1974) J. Virol. 13, 428-438. 31. Ruckman, J., Parma, D., Tuerk, C., Hall, D. H. & Gold, L. (1989) New Biol. 1, 54-65. 32. Varani, G., Cheong, C. & Tinoco, I., Jr. (1991) Biochemistry 30, 3280-3289. 33. Adari, H. Y., Rose, K., Williams, K. R., Konigsberg, W. H., Lin, T.-C. & Spicer, E. K. (1985) Proc. Natl. Acad. Sci. USA 82, 1901-1905. 34. Gollnick, P., Ishino, S., Kuroda, M. I., Henner, D. J. & Yanofsky, C. (1990) Proc. NatI. Acad. Sci. USA 87, 8726-8730. 35. Pabo, C. 0. & Sauer, R. T. (1984) Annu. Rev. Biochem. 53, 293-321. 36. Ronson, C. W., Nixon, B. T. & Ausubel, F. M. (1987) Cell 49, 579-581. 37. Jiang, J., Gu, B., Albright, L. M. & Nixon, B. T. (1989) J. Bacteriol. 171, 5244-5253. 38. Chou, P. Y. & Fasman, G. (1978) Adv. Enzymol. 47, 45-148. 39. Michel, B. & Zinder, N. D. (1989) Proc. Natl. Acad. Sci. USA 86, 4002-4006. 40. Houman, F., Diaz-Torres, M. R. & Wright, A. (1990) Cell 62, 1153-1163. 41. Miller, E. S., Winter, R. B., Campbell, K. M., Power, S. D. & Gold, L. (1985) J. Biol. Chem' 260, 13053-13059.
