of three identical 71,600-dalton chains, the products of gene. 9 (14-16). During the maturation of the native trimer, newly synthesized polypeptide chains partially ...
Proc. Natl. Acad. Sci. USA Vol. 81, pp. 6584-6588, November 1984 Biochemistry
Single amino acid substitutions influencing the folding pathway of the phage P22 tail spike endorhamnosidase (protein folding/amino acids/DNA sequences/temperature-sensitive mutants)
MYEONG-HEE YU AND JONATHAN KING Department of Biology, Massachusetts Institute of Technology, Cambridge, MA 02139
Communicated by Robert L. Baldwin, June 15, 1984
Temperature-sensitive mutations in the gene ABSTRACT for the thermostable tail spike of phage P22 interyFere with the folding and subunit association pathway at the restrictive temperature but not with the activity or stability of the protein once matured. The local sites of these mutations and the mutant amino acid substitutions have been determined by DNA sequencing. Of 11 temperature-sensitive folding mutations, 3 were replacements of glycine residues by polar residues, and three were replacements of threonine residues by residues unable to form a side-chain H-bond. There were no proline replacements. Two of the temperature-sensitive sites in which threonine residues were replaced by isoleucine residues were homologous. These sequences probably maintain the correct local folding pathway at higher temperatures. The temperature-sensitive amino acid substitutions appear to destabilize a thermolabile intermediate in the wild-type folding pathway or to increase the rate of a competing off-pathway reaction.
mediate, which is NaDodSO4-sensitive and trypsin-sensitive, then folds further, generating the NaDodSO4- and trypsin-resistant native tail spike (16). The protrimer is separable from the native spike, migrating more slowly during native gel electrophoresis (17). The crystal structure of the tail spike has not been determined. Raman spectroscopy studies indicate a predominance of P-sheet structure (18). ts mutations have been mapped to more than 35 sites in the tail spike gene (19), and 15 of them have been characterized in detail (10-12). At the 30'C permissive temperature, the mutant polypeptide chains form the native spike. At the 40'C restrictive temperature, the polypeptide chains are synthesized, but no active tail spikes are formed. This lack of native protein is not due to thermolability of the ts mutant proteins. The mutant proteins formed under permissive conditions are as thermostable as the wild-type protein, requiring heating to 90'C for inactivation (10). A variety of criteria, including sensitivity to trypsin digestion, sensitivity to detergent, and antigenicity, show that the inactive chains synthesized at high temperature are in a different conformation than the wild-type spike at high temperature or the mutant spikes at low temperature (11, 12). Goldenberg et al. (12) showed that the ts polypeptide chains were blocked at stages in the pathway prior to the partially folded protrimer intermediate. Upon cooling to permissive temperature, the accumulated ts chains formed the protrimer and continued through the pathway to the mature spike (11, 12). Thus, the ts mutant polypeptides accumulating at restrictive temperature are either earlier intermediates in the folding pathway or represent off pathway states that nonetheless are related reversibly to true intermediates. These mutations correspond to that class referred to as TSS-Temperature-Sensitive Synthesis-by Sadler and Novick (20). Since the mutations do not affect the synthesis of the gene 9 polypeptide chain, and since there is no evidence in the tail spike pathway for native monomers or for any step corresponding to the assembly of native monomers, we will refer to these mutants as ts folding-TSF-mutants rather than TSS mutants. Matthews et al. (21) have carefully studied a missense mutation in tryptophan synthetase, Gly211 Glu, which does not prevent chain folding but does alter both the unfolding and refolding pathways. The effects were complex but were most marked in the early phase of refolding. The complete sequence of gene 9 has been determined by Sauer et al. (22). In the present work, we precisely located the sites of 12 ts folding mutations in the gene and polypeptide chain and identified the mutant amino acid substitutions. The results are interpreted in terms of the character of the code through which amino acid sequence determines protein conformation.
A fundamental aspect of gene expression is the folding of newly synthesized polypeptide chains into correctly ordered three-dimensional structures. Since the structure of many small proteins is fully determined by their amino acid sequences, features of the sequence must direct the folding of the polypeptide chain (1, 2). However, the code that relates amino acid sequences to three-dimensional structures remains undeciphered (2, 3). Evidence from a number of studies suggests that some residues are not important in the folding process, whereas others are critical. Comparison of homologous proteins from different organisms show clearly that many residues in a sequence can be altered and still yield the same final native structures (4-7). On the other hand, studies with, for example, the NH2-terminal peptide of ribonuclease have identified amino acid residues critical for correct folding and conformation (8, 9). We have adopted a genetic approach to identifying in a polypeptide chain those residues and local sequences that influence the folding pathway. This involves characterization of temperature-sensitive (ts) point mutations which block the formation of an intermediate in the maturation of a structural protein, the tail spike endorhamnosidase of Salmonella typhimurium phage P22 (10-12). The tail spikes of phage P22 are thermostable proteins that participate in the attachment of the phage to the host cell. The protein has an endorhamnosidase activity that cleaves the O-antigen of Salmonella (13). The spikes are composed of three identical 71,600-dalton chains, the products of gene 9 (14-16). During the maturation of the native trimer, newly synthesized polypeptide chains partially fold and then associate into the protrimer, an intermediate in which the three chains are associated but not fully folded (Fig. 1). This interThe 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.
Abbreviation: ts, temperature-sensitive.
6584
Biochemistry:
Yu and
King
~ ~ ~I
Proc. NatL Acad Sci. USA 81 (1984)
6585
ts Defects
c
\
L
Folded Chains _J
/
~~~~~~~~~~~~~~I
(AggregatedA States /, 40 OC-'30°C t
II"w
Translation 4
I
//~~~~~~~~-/ k
=[
P
artia
Chains
[I
ix
Mature Mutant
Spike Stable up to
80C
1
%][Protrimeri Chains Associated but not fully folded
FIG. 1. Tail spike maturation pathway and the effects of ts mutations (12, 17). Wild-type tail spikes form at both 30'C and 40'C but with reduced efficiency at 40'C. The ts mutations sharply reduce the high temperature yield but have only slight affects on the pathway at low temperature. Once the mature spike is formed, wild-type and ts mutant proteins are stable well above 40TC. The ts stage in the wild-type pathway is before the protrimer, either the partially folded monomeric intermediate, or its formation. The inactive polypeptide chains accumulating at high temperature are found after cell lysis as large aggregates (unpublished data).
MATERIALS AND METHODS Strains. S. typhimurium DB7000 (su-) was used as a host for all experiments. DB7004 (leuAam414;supE) was used as the permissive strain for the growth of phage carrying amber mutations (23). All of the phage strains used carried a mutation in the cI gene (cI-7) to ensure entry into the lytic cycle. The ts mutations in gene 9 have been described previously (11). Phage strains used for DNA preparation carried in addition to the ts mutation an amber mutation in gene 13 (amH101), which delays cell lysis. Strains for pulse-chase labeling experiments carried an additional mutation in gene 5 (amN114), which specifies the coat protein. In the absence of coat protein, the assembly of head structures is blocked, and the tail spike protein remains as a major soluble protein inside the infected cells. Determination of Nucleotide Substitutions and Amino Acid Substitutions. The gene for the tail spike protein has been sequenced by Sauer et al. (22). DNA from mutant phage was prepared by phenol extraction from two cycles of CsCl gradient-purified phages. Digestion of P22 chromosome with EcoRI and BamHI restriction enzymes yielded several fragments. They were separated by electrophoresis on a lowmelting-point agarose gel. The fragment containing most of the tail spike gene (1.8 kilobase pairs) was eluted from the gel by phenol extraction. From the genetic map and the size of various amber fragments, each ts mutation could be approximately located on a smaller restriction fragment of gene 9. In addition, the end points of deletion lysogens had been determined by restriction mapping (P. Berget, University of Texas, Houston, personal communication), against which these ts mutations in gene 9 were originally mapped. This allowed us to locate each mutation more precisely. Appropriate restriction fragments of mutant DNA were end-labeled, isolated, and sequenced by the method of Maxam and Gilbert (24). The amino acid substitutions were deduced from the nucleotide substitutions. Pulse-Chase Labeling Protocol. Pulse-chase labeling experiments were performed as described (17) with some modification. Exponentially growing DB7000 culture (2 x 108 cells per ml) in M9 minimal medium at 30'C was infected with phage strains at a multiplicity of infection of 5. After 1 hr, infected cells were labeled with '4C-labeled amino acids (3 ,Ci per ml of culture; 1 Ci = 37 GBq) for 1 min and chased with 10% acid hydrolyzed-casein (60 Al per ml of culture) at 300C. At 1 min after chase, portions of a culture were transferred to six different temperatures between 15'C and 370C. At various times afterwards, samples were withdrawn, lysed
by freezing and thawing, and then analyzed by gel electrophoresis. RESULTS Determination of the Amino Acid Substitutions Causing ts Defects. Fourteen of the 15 gene 9 ts mutations studied by Goldenberg et al. (12) interfered with the spike maturation pathway. Twelve of these were chosen for sequencing, as well as an independent ts mutation at the same site as 1 of the 12. The size of the tail spike gene is close to 2000 base pairs (666 amino acid residues), so that sequencing the whole gene for each mutant was not practical (22). With use of the fine structure map of gene 9 previously determined (19) and knowledge of the deletion end points (P. Berget, personal communication), each ts mutation was roughly located on the physical map. The restriction fragment most likely to contain the ts mutation was isolated from each mutant strain. Fig. 2 shows the fragments isolated and the regions sequenced for the mutants. Restriction fragments were chosen in such a way that only a single end was labeled by the Klenow fragment with either [a-32P]dATP or [a-32P]dTTP at BstNI sites and with [a-32P]dCTP at Acc I sites. For most of the mutations shown in Fig. 2, the regions of the restriction fragments that were sequenced displayed a single nucleotide change in comparison with the corresponding wild-type sequences (Fig. 3). For tsH304, the restriction digest pattern indicated the loss of the BstNI site at nucleotide 729. The Acc I fragment containing the BstNI site was isolated and sequenced for this mutation (Fig. 2). For tsU56, no substitution was found throughout the entire middle BstNI fragment (729-1227). For tsU11, only one fragment, Msp I(573)-BstNI(729) was sequenced, where no substitution was found. These latter two mutations remain to be located and sequenced. The nucleotide substitutions and the corresponding amino acid replacements are shown in Table 1. Two mutations, tsRH and tsRAF, which mapped at the same site in two factor crosses, had the same nucleotide substitutions. Eight of the 11 mutations sequenced were due to transitions, 7 of them being G-C to A-T. The remaining three of the mutations were due to transversions. These are consistent with previous findings for the distribution of UV and hydroxylamine mutations and the effects of error prone repair (25, 26). Among the 11 amino acid residues that were replaced, 3 were glycines and 3 were threonines, with single substitu-
6586
Biochemistry: Yu and King Msp I
573
li
~-
8s f NI AccI Acc I 729 779
1
879
,
Proc. NatL Acad Sci. USA 81 8st NI 1227
1.-A-
amH 200, U5,H300, U t
(1984)
8st NI 1457
.1
9.1, H301, U38, amEI017
U24, RAF, RH, U19, H302, U56, amH840 amH840, H302, H304, U18, U56, 9.1
H 304
H301, U18,U56, 9.1
FIG. 2. Regions of the tail spike gene sequenced for ts mutants. The bar represents a portion of the gene. Numbers indicate the nucleotide in base pairs starting from the initiating codon. Restriction sites used for sequencing are indicated by short vertical lines. Circles indicate positions of 3-end labeling. Arrows indicate extent of reliable sequencing.
toration of the BstNI site (729) was examined. All revertants but that from ts9.1 regained the wild-type sequences at the altered sites. This confirmed that those amino acid substitutions are responsible for the ts folding phenotype of the mutant tail spike proteins. The revertant isolated from ts9.1 still retained the substitution, suggesting that it might not represent the ts mutation. However, the site of the nucleotide substitution corresponded to the position of the ts mutation on the genetic map. We suspect that the substitution found is the ts mutation and that the revertant has a second site mutation that suppresses the original ts mutation. Temperature Dependence of Tail Spike Maturation. As a means of probing the kinds of interactions that the residues at the mutant site were engaged in, we examined the kinetics of tail spike maturation as a function of temperature. Infected cells were exposed to a short pulse of amino acids, and then portions were transferred to different temperatures and incubated further. Samples were taken at various times, and the reaction was stopped by freezing. After thawing and lysis in the cold, samples were analyzed by NaDodSO4 gel electrophoresis without prior heating above room temperature. Under these conditions the mature tail spike is resistant to NaDodSO4 dissociation and can be distinguished from the nonnative precursors, which migrate together as an NaDodSO4-polypeptide complex (16). Fig. 4 shows the tail spike maturation kinetics for wild type and tsH300 (Thr235 Ile). The abscissa shows the native spike formed as a percentage of the total pool of newly synthesized tail spike polypeptide chains. Fig. 4A shows the formation of native tail spikes from the wild-type polypeptide chains at six different temperatures. The solid lines are the higher temperature data, while the dotted lines show
tions of valine, alanine, isoleucine, arginine, and serine. None of the substituted residues was proline. Statistical analysis showed that the observed frequency of the substitutions did not differ significantly from that expected at random except for the threonine substitutions. The mole percentage of threonine in the region is 7.8%. The occurrence of ts mutations at threonine 3 times out of 11 is 2.4 standard deviations more frequent than expected assuming a random distribution, significant at the 99.5% level. The mutant amino acid substitutions were quite diverse, including more hydrophobic Ser227 Phe, more hydrophilic Gly2 -4 Arg, and neutral Ile258 -* Leu.Most of the local sequences in which the mutations fell did not reveal any obvious homologies or patterns. However, the two local sequences containing the threonine-to-isoleucine replacements-tsH300 and tsH301-have the same amino acids at three positions and related residues at two positions at the mutation sites (Table 1). The NH2-terminal residues to the threonine residues are glycine and proline, both helix breakers and common at reverse turns. The tryptophan and tyrosine residues NH2-terminal to the glutamine residues are also related residues. Since the same replacement in both sequences generated similar ts defects, the sequences probably affect related local conformations during polypeptide chain maturation. Analysis of Revertants. To determine if the nucleotide substitutions represented the ts mutations, spontaneous revertants were isolated from a subset of the mutants. Plaques formed at high temperature were isolated from mutants tsH304, tsRH, tsU24, tsUJ9, tsUI8, and ts9.1. Stocks of the revertant phage were grown up from single plaques, and the same fragment containing the original substitution was isolated and sequenced. For the revertant from tsH304, the res_,
-
0
"M
lot% %
da oft
I* fj0 %-a
wo a&
4% oft --AV
4
.1q 0
% .k
A>C G H302 RH U19 U24 H302 RH U19
G A>C C+T C
U241
WF
C C+ T H302 RH U19 U24 H302 RH U19 U24
FIG. 3. DNA sequencing gel showing the tsU24 substitution. The EcoRI/BamHI DNA fragments containing gene 9 were digested with BstNI, end-labeled with [a-32P]dTTP, and the appropriate subfragments were isolated by polyacrylamide gel electrophoresis. DNA sequencing was performed by the method of Maxam and Gilbert (24). In the case of tsU24, the nucleotide change of TAT -- AAT shown in the figure (arrows) corresponds to the amino acid substitution of Leu for Ile at residue position 258. WT, wild type.
Biochemistry:
Yu and
King
Proc. Natl. Acad. Sci. USA 81 (1984)
Table 1. Nucleotide and amino acid substitutions of gene 9 ts folding mutants Amino acid Mutation Codon change Residue ++ mutation TCT--TTT tsU5 227 Ser - Phe Thr Leu Lys tsH300 ACC-*ATC 235 Thr Ile Gly Tyr Gln tsH304 GGA-+AGA 244 Val Lys Phe Gly Arg tsU24 ATA-KITTA 258 Ile -. Leu Lys Gly Gln tsRAF GTC- GGC 270 Val Gly Glu Cys Ile tsRH GTC-*GGC 270 Val - Gly Glu Cys Ile tsUI9 AGA--AAA 285 Arg - Lys Gly Phe Leu tsU18 ACC--GCC 307 Thr Ala Asp Gly Ile tsH302 GGC--GTC 323 Asn Tyr Val Gly Asp ts9.J GCC-*GTC 334 Ala Val Ser Val Ser tsH301 ACT-*ATT 368 Thr Ile Thr Trp Gln tsU38 GGG--GAG 435 Leu Leu Val Gly Glu
maturation in the lower temperature range. At lower temperatures the overall rate of maturation slowed down, while the final yield was not affected. In the higher temperature range exhibited by solid lines in Fig. 4A, the overall rate of maturation slightly increased, but the yield of native tail spikes decreased. The yield of mature tail spikes at 370C was about 40% of the yield at 250C. Fig. 4B shows the maturation kinetics of tsH300 (Thr235 Ile). At temperatures below 250C (broken lines), the curves are almost identical to those of wild type. However, at 30'C the yield of spikes decreased to "50% of the wild-type yield, and at higher temperatures it was reduced even further. The rates of maturation did not appear to be affected.
E 6( ._
c4,
250 200 150
300 330 370
10
20 30 40 5 Time after chase (min)
FIG. 4. Temperature dependence of in vivo formation of mature tail spikes. Bacteria were infected at 30°C with either 9+/5-/13-/cI (A) or tsH300/5-/13-/cI (B). After 60 min, 14C-labeled amino acids were added to the infected cells and then chased for 1 min with cold amino acids. One minute after chase, portions of culture were transferred to 15°C (A), 200C (X), 250C (o), 300C (A), 330C (m), and 37°C (W). At various times afterwards, samples were withdrawn and lysed by freezing and thawing. The samples were electrophoresed through NaDodSO4/polyacrylamide gels without prior heating. The peaks corresponding to the mature trimer and NaDodSO4/polypeptide chain complex were quantitated. The abscissa represents gp9 in mature spikes per total gp9.
Local sequences Gln Ser Lys Thr Asp Pro Thr Val Ser Asp Pro Gly Ile Glu Thr Asn Ile Thr Ser Thr Gly Val Glu Val His Gly Val Glu Val His Phe Arg Gly Cys His Ile Thr Phe Glu Asn Ile Gly Gly Arg Thr Ser Ala Gln Phe Leu Gly Thr Val Gly Ser Arg Gly Ala Leu Gly
6587
Gly Tyr Leu Leu
Arg Arg Phe Leu Ser
Arg Thr Val
Similar data on the kinetics and yield of tail spike formation as a function of temperature were collected for the polypeptide chains from tsH301 (Thr368 - Ile), tsH304 (Gly44 -* Arg), and tsU24 (Ile258 -* Leu). In all cases the final yields declined with increasing temperature more sharply than with wild type, while the initial rates were not affected. The profiles of yield versus temperature for tsH304 and tsU24 were similar. The profiles for tsH301 and tsH300 (Thr235 _+ Ile) were also similar to each other but were distinguishable from tsH304 and tsU24, with steeper slopes.
DISCUSSION Previous studies have revealed that the maturation of the wild-type tail spike is itself sensitive to temperature, with
