Protein Science Sheba protein
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
QUICK SEARCH:   [advanced]


     

This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow Alert me when this article is cited
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow reprints & permissions
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Gage, M. J.
Right arrow Articles by Robinson, A. S.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Gage, M. J.
Right arrow Articles by Robinson, A. S.
Social Bookmarking
Add to CiteULike   Add to Connotea   Add to Del.icio.us   Add to Digg   Add to Reddit   Add to Technorati  
What's this?
Protein Science (2003), 12:2732-2747.
Copyright © 2003 The Protein Society

C-terminal hydrophobic interactions play a critical role in oligomeric assembly of the P22 tailspike trimer

Matthew J. Gage and Anne Skaja Robinson

Department of Chemical Engineering, University of Delaware, Newark, Delaware 19716, USA

Reprint requests to: Anne Robinson, 259 Colburn Laboratory, Department of Chemical Engineering, University of Delaware, Newark, DE 19716, USA; e-mail: robinson{at}che.udel.edu; fax: (302) 831-6262.

(RECEIVED April 21, 2003; FINAL REVISION September 8, 2003; ACCEPTED September 8, 2003)

Article and publication are at http://www.proteinscience.org/cgi/doi/10.1110/ps.03150303.


   Abstract
TOP
 Abstract
Introduction
 Results
 Discussion
 Materials and methods
 References
 
The tailspike protein from the bacteriophage P22 is a well characterized model system for folding and assembly of multimeric proteins. Folding intermediates from both the in vivo and in vitro pathways have been identified, and both the initial folding steps and the protrimer-to-trimer transition have been well studied. In contrast, there has been little experimental evidence to describe the assembly of the protrimer. Previous results indicated that the C terminus plays a critical role in the overall stability of the P22 tailspike protein. Here, we present evidence that the C terminus is also the critical assembly point for trimer assembly. Three truncations of the full-length tailspike protein, TSP{Delta}N, TSP{Delta}C, and TSP{Delta}NC, were generated and tested for their ability to form mixed trimer species. TSP{Delta}N forms mixed trimers with full-length P22 tailspike, but TSP{Delta}C and TSP{Delta}NC are incapable of forming similar mixed trimer species. In addition, mutations in the hydrophobic core of the C terminus were unable to form trimer in vivo. Finally, the hydrophobic-binding dye ANS inhibits the formation of trimer by inhibiting progression through the folding pathway. Taken together, these results suggest that hydrophobic interactions between C-terminal regions of P22 tailspike monomers play a critical role in the assembly of the P22 tailspike trimer.

Keywords: P22 tailspike; assembly; ANS; folding; hydrophobic interactions


   Introduction
TOP
 Abstract
 Introduction
Results
 Discussion
 Materials and methods
 References
 
Protein–protein interactions are of fundamental importance to the proper function of a cell. Examples of common protein–protein interactions include antibody–antigen interaction (Smith et al. 1993), signal transduction pathway regulation (Onofri et al. 2000), transcription regulation (Buratowski et al. 2002), chaperone binding (Kang et al. 1994), and enzyme regulation (Smith et al. 2002). Protein complexes range in size from dimers (Gloss and Matthews 1998) to large heterocomplexes (Jager and Pata 1999; Reinisch et al. 2000).

Correct formation of subunit interactions is critical to the proper function of the protein complex. Some proteins, such as the {alpha} and ß proteins in the bacterial tryptophan synthase complex, are capable of functioning as monomers (Miles et al. 1987; Miles 1991). However, when combined to form an {alpha}2ß2 complex, the enzymatic activity of the protein is enhanced by up to two orders of magnitude (Miles 1995). In other cases, the formation of intersubunit contacts is critical for activity, such as with bacterial luciferase, where the protein becomes active only after the quaternary structure of the protein is fully formed (Baldwin et al. 1993).

Subunit contacts can be as simple as domain–domain contacts, such as in the NAD-dependent dehydrogenases (Jaenicke 1987), where the subunits are held together by contacts along a distinct subunit interface. Subunit contacts also include more complex interactions, such as in coiled-coil domains, where the subunit becomes intertwined and the native structure is dependent on interaction between subunits during folding (Ban et al. 2000; Lehman et al. 2001). A critical point in the assembly of complex protein interactions is the initial nucleation step. Proper interdigitation of the collagen polypeptide chains begins with an initial nucleation event involving the C terminus (Buevich and Baum 2001). The remainder of the protein can be properly folded once the nucleus is properly constructed.

Folding of the polypeptide chain and assembly of the protein complex are coupled processes in many multimeric proteins. Examples where the native structure can only be reached through coupled folding and assembly include coiled-coils (Burkhard et al. 2001), collagen (Bachinger et al. 1980), multisubunit ß-helices (Kreisberg et al. 2000), the P22 Arc repressor dimer (Milla et al. 1995), and the adenovirus tail fiber (Van Raaij et al. 1999). In these proteins, the stability of the monomer is dependent on contacts between the subunits, and stable structures are unable to form in the absence of critical intersubunit contacts. The same forces that stabilize monomeric proteins also stabilize protein–protein interfaces, namely hydrogen bonds, hydrophobic interactions, and ionic interactions (Hiraga and Yutani 1997; Labrie et al. 2001; Burkhard et al. 2002).

One protein with coupled folding and assembly processes is the tailspike protein from the Salmonella typhimurium phage P22. P22 tailspike is a homotrimer composed of 666 amino acid polypeptide chains (Fig. 1Go). The mature P22 tailspike protein is both SDS- and protease-resistant (Goldenberg and King 1981). This resistance is not acquired until the protein is completely matured, which has aided in the study of the folding pathway. P22 tailspike is one of the few proteins for which both in vivo and in vitro folding intermediates have been identified (Goldenberg and King 1982; Danner et al. 1993).



View larger version (64K):
[in this window]
[in a new window]
  		Figure 1. Structure of the P22 tailspike protein. (A) Ribbon diagram of the structure of P22 tailspike protein. The three subunits are colored red, blue, and yellow. The structure was solved in two parts: the main body in 1994 (Steinbacher et al. 1994) and the head binding domain in 1997 (Steinbacher et al. 1997). The four main structural features are indicated on the structure. The arrows highlighting residues 109 and 544 indicate the approximate location of the truncations made in the present study. (B) Ribbon diagram of the ß-helix domain interface. This view is looking down the interface between the three ß-helix domains from the N terminus. The A, B, and C ß-sheets that compose the helix are indicated. The three ß-helices that compose the helix are colored as in panel A. The light-blue balls in the space between the subunits indicate the position of the ordered waters in the structure. Note that the interactions between the domains appear to be predominately mediated by water. (C) Ribbon diagram of the C terminus, oriented to view down the long axis of the protein from the ß-helices to the end of the protein. The D and E ß-sheets that compose the C terminus are indicated, and the three subunits are colored as in panel A. The positions of the four amino acids mutated are shown in ball-and-stick representation. (D) Surface charge diagram of a monomer of the P22 tailspike protein generated using the program GRASP (Nicholls et al. 1991). The protein is in the same orientation as the blue subunit in panel A. Blue, positively charged regions; red, negatively charged regions; white, hydrophobic regions. Arrows highlight the two hydrophobic patches on the C terminus.

 
P22 tailspike consists of three main structural elements: the head-binding domain at the N terminus, the ß-helix in the center of the protein, and the ß-prism and caudal fin at the C terminus (Fig. 1AGo). In the head-binding region, each monomer forms two ß-sheets, which lie perpendicular to each other and are stabilized by a hydrophobic core (Steinbacher et al. 1997). In this region, there are very few intersubunit contacts. The main body of the protein is formed by three right-handed parallel ß-helices, one from each monomer, which lie parallel to the long axis of the protein (Steinbacher et al. 1994). Each helix is composed of three 13-stranded ß-sheets: A, B, and C. ß-sheets A and B form the subunit interface, which is dominated by hydrophilic groups and ordered water molecules (Fig. 1BGo). The ß-prism and caudal fin domains, shown in Figure 1CGo, are formed by ß-sheets D and E from the three monomer chains to generate a prism-like structure (Steinbacher et al. 1994). The center of this prism is filled with nonpolar residues, and there are no ordered waters in the core of this region of the protein. An electrostatic potential map of residues 109–666 from one monomer shows the existence of a large, hydrophobic patch in the C terminus that is exposed in the monomer but which becomes buried in the trimer (Fig. 1DGo).

The folding pathway of P22 tailspike has been characterized both in vivo and in vitro Goldenberg and King 1982; Goldenberg et al. 1983; King and Yu 1986; Haase-Pettingell and King 1988; Seckler et al. 1989; Danner and Seckler 1993; Danner et al. 1993; Beibinger et al. 1995). Folding occurs rapidly (on the order of minutes) in vivo and on the order of hours, depending on temperature, in vitro (Goldenberg and King 1982; Danner et al. 1993). Figure 2Go is a schematic of the current model of the in vitro folding and aggregation pathways. The initial step is the folding of the unfolded polypeptide chain to form a stable monomer species. Seckler et al. showed that this initial folding step is predominantly the formation of the ß-helix domain (Miller et al. 1998). Partially folded monomer chains add to one another to form dimer and subsequently a protrimer species. The protrimer species undergoes currently undetermined structural rearrangements to generate the mature tailspike trimer. One interesting feature of the folding of P22 tailspike is the formation of a transient disulfide bond during folding (Robinson and King 1997; Haase-Pettingell et al. 2000). There are eight cysteine residues in P22 tailspike. In the final, mature form of the protein, all eight cysteines are reduced (Sargent et al. 1988; Steinbacher et al. 1994). However, single-mutant substitution of the three C-terminal cysteines has been shown to significantly affect folding and/or assembly kinetics in vivo (Haase-Pettingell et al. 2000) and in vitro (Danek and Robinson 2003).



View larger version (17K):
[in this window]
[in a new window]
  		Figure 2. In vitro folding and aggregation pathway of P22 tailspike. Unfolded protein begins to fold and enters either the folding pathway (IM) or the aggregation pathway (IM*). The monomer species may change between folding-competent and aggregate-prone conformations depending on conditions until it dimerizes with another monomer in the same state. Interaction between the pathways in the dimer state has also been proposed. The folding-competent dimer adds a monomer to make the protrimer species that then undergoes structural rearrangement to form the final trimer species. Aggregates add monomers sequentially and also as clusters to form large aggregates.

 
A percentage of the tailspike polypeptide chains form insoluble aggregates, even under the most ideal conditions in vivo. It has been shown that the GroE chaperones interact with partially folded tailspike molecules, though interaction with the chaperones does not actually facilitate assembly (Brunschier et al. 1993). Instead, addition of GroEL during refolding traps refolding intermediates, blocking trimer formation (Brunschier et al. 1993). Also, addition of both GroEL and GroES together does not improve refolding yields. Clark and King (2001) demonstrated that nascent polypeptide chains remain associated with ribosomes for extended periods of time following translation. They proposed that the ribosome might act as a nontraditional chaperone and interact with the nascent polypeptide chain to prevent unfavorable associations. Both of these results suggest a possible role for hydrophobic interactions in assembly of the protrimer.

The C terminus was proposed as the possible assembly point of the protrimer when the structure of the 109–666 region of the tailspike protein was determined (Steinbacher et al. 1994). Our present study provides evidence to support the hypothesis that oligomerization is C-terminally driven. First, an intact C terminus was shown to be essential for stable trimer formation. Truncations of the P22 tailspike protein that do not contain the C terminus exist in a reversible equilibrium between monomer and trimer, strongly favoring the monomeric form. In addition, C-terminal truncations were unable to form mixed species with the full-length P22 tailspike protein. Second, introduction of charged and polar residues into the hydrophobic core of the C terminus generally blocks trimer formation, suggesting that the hydrophobic core of the ß-prism domain and the caudal fin provides the driving force for assembly. Finally, addition of ANS, a probe for hydrophobic patches on proteins, during refolding inhibited trimer formation by blocking the progression through the folding pathway. These results support the hypothesis that the hydrophobic regions of the C terminus provide the driving force for oligomerization.


   Results
TOP
 Abstract
 Introduction
 Results
Discussion
 Materials and methods
 References
 
Characterization of the TSP{Delta}C truncation
The C terminus has long been thought to play a role in assembly of the trimer. Steinbacher et al. (1994) suggested that the intersubunit contacts in the C terminus likely played a critical role in the formation of the protrimer. More recently, work by Seckler and colleagues on the ß-helix domain (residues 109–544) demonstrated that the protrimer species must be stabilized by contacts besides those between ß-helix domains (Miller et al. 1998), and work from King and coworkers has shown that mutation of G546 to aspartic acid permitted protrimer formation while preventing formation of the mature trimer, further implying a role for the C terminus in trimer formation (Kreisberg et al. 2002). These results support the hypothesis that the C terminus is critical for trimer formation. The question is, what are the critical interactions in the C terminus that drive formation of the trimer?

To begin to answer this question, three truncations, TSP{Delta}N, TSP{Delta}NC, TSP{Delta}C, were generated to investigate the role of the three main domains of the protein; the head-binding domain, the ß-helix domain, and the C-terminal domain. The TSP{Delta}N truncation, consisting of residues 109–666, was used to determine the original crystal structure. Previous work had demonstrated that the TSP{Delta}N truncation formed a stable trimer species with properties similar to the full-length trimer (Danner et al. 1993). Seckler and colleagues showed that this construct is capable of forming mixed trimers when combined with full-length P22 tailspike (Danner et al. 1993); therefore, it is a useful control for mixed assembly experiments with the other truncations.

The TSP{Delta}NC truncation, consisting of residues 109–544, has been used to study the formation and stability of the ß-helix domain (Miller et al. 1998). This construct was reported to form an equilibrium between monomer and trimer species, presumably through interactions similar to those in the mature full-length trimer (Miller et al. 1998). Higher protein concentrations favored trimer, whereas lower protein concentrations favored monomer, with a midpoint of ~0.5 mg/mL (Miller et al. 1998). The ability of this protein to assemble into mixed species with the full-length protein has not been determined.

The final truncation, TSP{Delta}C, consists of residues 1–544 and has not been previously characterized. Spectroscopic characterization of the conformation of the TSP{Delta}C was performed by both circular dichroism (CD) and fluorescence. Fluorescence emission was measured by excitation of the tryptophan residues at 280 and measuring the emission spectrum from 300 to 400 nm. The spectrum of TSP{Delta}C had a peak intensity of 340 nm and a center of mass of 342 nm, identical to that of both the full-length tailspike protein and the TSP{Delta}NC truncation (Fig. 3AGo). In addition, the emission spectra exhibited a shape similar to the spectra of the full-length protein. The CD spectrum was measured from 260 to 190 nm and showed a single minimum at 220 nm (Fig. 3BGo). This is red-shifted about 4 nm from the value reported in the literature for both the TSP and TSP{Delta}NC proteins. However, the minimum of the peak measured for TSP{Delta}C was identical to the minimum that we measured for both the full-length tailspike protein and the TSP{Delta}NC under identical conditions, and therefore the shift is believed to be due to the buffer conditions and not due to a difference in the protein structure. The similar shape and characteristics of the spectra indicate that the molecular environment of the aromatic residues is very similar and also that the backbone exists in a similar conformation.



View larger version (16K):
[in this window]
[in a new window]
  		Figure 3. TSP{Delta}C has spectroscopic characteristics similar to those of the native tailspike trimer. Fluorescence emission (A) and CD (B) spectra of the TSP{Delta}C truncation. Spectra were recorded using 100 µg/mL TSP{Delta}C diluted in 1X (fluorescence) or 0.01X (CD) Tris Refolding Buffer. The fluorescence spectrum has a center of mass of 342 nm, which is identical to the center of mass of the native tailspike. The CD spectrum has a minimum at 220 nm, consistent with native tailspike protein spectra collected on the same instrument.

 
Sedimentation equilibrium experiments were performed to determine whether the TSP{Delta}C truncation existed in solution as a monomer or as an oligomeric species. It has been reported that the TSP{Delta}NC truncation exists in an equilibrium between monomer and trimer with an association constant of 7 x 109 M-2 (Miller et al. 1998). Concentrations of 2 µM and 6 µM TSP{Delta}N were sedimented at speeds of 10,000, 12,500, and 15,000 rpm, and the data were fit both to a single monomeric species and to a self-association model. The fits for each model and the resulting residuals for the 15,000 rpm data are shown in Figure 4Go. Fitting the data to a single monomeric species resulted in molecular weight predictions of 60.7 kD and 68.8 kD for the 2 µM and 6 µM solutions respectively, both of which are larger than the actual molecular weight of the protein. This increase in the apparent molecular weight is consistent with a protein that has a weak concentration-dependent self-association. In addition, the residuals from fitting the 6 µM solution to a monomeric species indicate that a more complicated model is necessary to properly fit the data.



View larger version (28K):
[in this window]
[in a new window]
  		Figure 4. TSP{Delta}C exists as an equilibrium between monomer and trimer in solution. Analytical ultracentrifugation was performed on 2 µM and 6 µM solutions of TSP{Delta}C at 10,000, 12,500, and 15,000 rpm. Shown are the raw data at 15,000 K and the residuals for the fit of the data to a nonassociating monomer model (A) and to a self-associating monomer-trimer equilibrium model (B). As can be seen by the residuals, the 2 µM TSP{Delta}C data fit well to a single monomeric species model, whereas the 6 µM TSP{Delta}N data fit better to an equilibrium model between monomer and trimer species.

 
The 6 µM data was fit to a A|ZunA model, which produced an n value of 3.1 with improved residuals over the monomeric fit; the residuals for the 2 µM data did not improve by fitting the data to a self-association model (Fig. 4Go). This suggests that (1) the TSP{Delta}C truncation exists in a monomer-trimer equilibrium similar to the TSP{Delta}NC truncation, and (2) the protein exists primarily as a monomer at 2 µM. An association constant of 1.2 x 109 M-2 was calculated from the 6 µM equilibrium fit, corresponding to a free energy of association of -16.1 kJ/mole. These values are very similar to the values reported in the literature for the association of the TSP{Delta}NC domain, indicating that both the TSP{Delta}C and TSP{Delta}NC truncations exhibit a weak equilibrium between monomer and trimer, heavily favoring the monomeric species. This weak interaction between the subunits in the C-terminal deletion variants may play a role in assembly of the trimer.

Truncations missing ß-prism domain are unable to form mixed trimers
The role of each of the domains in trimerization was further investigated by mixing the three different truncations with full-length tailspike protein in in vitro refolding studies. All three truncations might be capable of forming stable trimer species with the full-length tailspike protein if the interactions along the ß-helix domain contribute significantly to trimerization. In contrast, only the TSP{Delta}N truncation should be able to form mixed species if the C terminus contains the critical elements required for association. In vitro refolding experiments were conducted with 50 µg/mL full-length P22 tailspike and equal molar ratios of the various truncations. TSP{Delta}N forms mixed species with the full-length P22 tailspike protein (Fig. 5A,BGo). In contrast, neither the TSP{Delta}NC truncation nor the TSP{Delta}C truncation was able to form mixed species with the full-length tailspike protein. The TSP{Delta}NC and TSP{Delta}C truncations were also unable to form mixed species in mixed refolding experiments with the TSP{Delta}N truncation (data not shown). These results support the hypothesis that the C terminus has a critical role in the trimerization process.



View larger version (41K):
[in this window]
[in a new window]
  		Figure 5. TSP{Delta}C and TSP{Delta}NC do not form mixed species with the full-length P22 tailspike protein. P22 tailspike was refolded in the presence of molar equivalents of three different truncations: TSP{Delta}N, TSP{Delta}C, and TSP{Delta}NC. Silver-stained and phosphor-images of nondenaturing PAGE gels using 14C-labeled full-length P22 tailspike are shown in A and B, respectively. Only the TSP{Delta}N truncation has the ability to form mixed species with the full-length tailspike protein. (C) Quantitation of the yields of full-length P22 tailspike in the presence of the truncations. Yields from three trials were averaged. As can be seen, no significant effect on yield was seen in the presence of most of the truncations. The decrease in trimer yield seen for 1 molar equivalent of 109–666 may be due to increased aggregation because of the high protein concentration. These results suggest that there is little or no effect on the yield of full-length tailspike when refolded in the presence of truncated versions of the protein.

 
Although TSP{Delta}N appears to be the only truncation capable of forming a stable trimer species with the full-length P22 tailspike, it is possible that either the TSP{Delta}NC or TSP{Delta}C truncation might induce aggregation of the full-length protein during folding and assembly, reducing the final yield of folded P22 tailspike protein. Yields of 14C-labeled full-length P22 tailspike were determined in the presence of varying molar ratios of the various truncations (Fig. 5BGo). Final trimer yields were calculated by integrating the area of trimer peak. Each experiment was carried out in triplicate, and the yields were averaged. The yields of folded full-length trimer in the presence of the truncations were within the error of the yield of full-length trimer for all of the experiments, with the exception of the TSP{Delta}N at an equal molar ratio (Fig. 5CGo). The decreased yield of trimer in the presence of an equal molar ratio of TSP{Delta}N is hypothesized to be due to increased aggregation due to high protein concentrations. Therefore, presence of the truncations did not significantly affect the yield of folded full-length P22 tailspike trimer.

Mutations in the hydrophobic core inhibit trimer formation
Because the C terminus contains the necessary elements for trimer formation, we explored the mechanism through which the C terminus promotes oligomerization. A study of the surface charge along the subunit interface shows the presence of a large hydrophobic region in the C terminus of the crystal structure (Fig. 1DGo). Burial of this region could potentially provide the driving force for oligomerization. If this hypothesis is correct, then disruption of the hydrophobic core by charged or polar residues should inhibit trimerization.

Four residues within the hydrophobic core of the ß-prism and caudal fin domains (L606, F636, A649, and L663) were chosen for mutagenesis. The position of the four amino acids within the C terminus is shown in Figure 1CGo. These positions were chosen because they are distributed throughout the C terminus and their side chains extended into the hydrophobic core of the C terminus. It appeared possible to alter the nature of the side chain of three of the four positions without significantly altering the size of the side chain. A series of mutations were generated using a combination of specific and degenerate primers, to examine the effects of polar residues, charged residues, and complimentary hydrophobic residues at each position (Table 1Go). It was predicted that polar and charged substitutions would disrupt trimer formation, whereas conservative hydrophobic substitutions would permit trimer formation.


View this table:
[in this window]
[in a new window]
  		Table 1. Effects of mutations in the C terminus on trimer formation
 
Each mutant gene was expressed in BL21(DE3) cells at 30°C for 4 h, under conditions identical to those used for expression of the wild-type P22 tailspike gene. Following expression, the cells were harvested by centrifugation and lysed. The soluble and insoluble fractions of the lysis were collected and analyzed using SDS-PAGE. A representative gel from this experiment is shown in Figure 6Go. A summary of the effects of all of the mutations on trimer formation is shown in Table 1Go.



View larger version (80K):
[in this window]
[in a new window]
  		Figure 6. Only conservative substitutions of residue 606 enable trimer formation in vivo. Protein expression was induced at 20°C using IPTG for 4 h. Samples were collected, lysed, and separated on 10% SDS-PAGE as described in Materials and Methods. Lane 1 contains wild-type TSP trimer and BSA as markers. Lanes 2,4,6,8,10,12, and 14 are of the supernatant (S) from the lysis. Lanes 3,5,7,9,11,13, and 15 were of the pellet (P) from the lysis. The individual mutants are indicated below the lanes in the gel. Lanes 2 and 3 show expression of the pET11a vector without the P22 tailspike gene. As can be seen, the L606V mutant is able to form trimer, whereas the substitution of charged or polar residues inhibits trimer formation.

 
Mutations at positions L606 and L663 exhibited the predicted trend. Substitution of either leucine with polar or charged residues resulted in the formation of insoluble monomers and no trimer formation. However, the three hydrophobic substitutions showed wild-type-like trimer formation. This is especially significant when one considers that L663 is only three residues from the C terminus of the protein, where the protein might be more capable of accommodating the introduction of a polar residue. However, introduction of a serine at position 663 completely abolished the formation of trimer, indicating that critical contacts exist throughout the entire hydrophobic core.

The mutations at F636 also showed a trend somewhat similar to that of the leucine mutations. Substitutions of the phenylalanine by polar or charged residues also inhibited trimer formation, except in the case of the tyrosine substitution, which formed trimer, though at a significantly lower level than wild-type tailspike. Tyrosine has a fairly hydrophobic nature, even with the addition of the hydroxyl group, and thus some limited trimer formation is not altogether surprising. The hydrophobic mutations also appear to inhibit trimer formation, in contrast to the leucine mutants. However, substitution of a proline for the phenylalanine is likely to introduce a kink in the ß-strand at F636, disrupting more than just the hydrophobic patch containing F636. Substitution of an alanine for phenylalanine is also likely to be destabilizing to the ß-strand, due to the large difference in size and decreased ß-sheet propensity. Therefore, the results from both the leucine and the phenylalanine mutants largely fit the model that the hydrophobic core of the protein plays a critical role in oligomerization.

The substitutions at A649 did not exhibit the predicted pattern. Both the serine and cysteine residues did not disrupt trimer formation as predicted, whereas the isoleucine and threonine substitutions were able to form trimer, though at a diminished capacity compared to wild-type tailspike. A close examination of the structure reveals that the A649 sites in the three chains are much further apart from each other than the other three residues examined (~12 Å for Ala vs. < 6 Å for Leu and Phe). This increased separation might account for the ability of the Cys and Ser mutants to form trimer. In addition, all of the substitutions have increased ß-sheet propensity, which might help compensate for the introduction of polar side chains. The decreased yields of trimer in the Thr and Ile mutants may be due to steric constraint, as both groups are much larger than the alanine they are replacing. The results for A649 indicate that proximity of the ends of the side chains is also critical, and that regions of the hydrophobic core that are not in close proximity contribute less to the potential driving force for trimerization.

Temperature-sensitive folding (tsf) mutations form native-like trimer but are incapable of forming trimer at higher temperatures. The effects of all of the known tsf mutants can be overcome by suppressor mutations at either V331 or A334. To verify that the mutations at L606 and L663 did not have a tsf phenotype, V331G mutations were made in the Leu->Asp and Leu->Lys mutations at positions 606 and 663. The double mutants were also incapable of forming trimer, indicating that the effects of the L606 and L663 mutations are not due to a tsf phenotype (Table 1Go).

The general trend exhibited by these substitutions is that polar and charged residues inhibit trimer formation, whereas hydrophobic substitutions allow trimer formation. This is consistent with the model that the hydrophobic core of the C terminus is providing the driving force for oligomeric assembly. The L663 mutations are particularly significant, as L663 is only three residues from the end of the protein. If the driving force for assembly was located outside the hydrophobic core of the C terminus, addition of either charged or polar residues at the C terminus might be readily compensated for by increasing the distance between the three polypeptide chains at the C terminus, allowing the formation of trimer.

ANS inhibits trimer formation
If the hydrophobic patches on the C terminus are causing the nucleation of trimerization, then a small molecule that binds to hydrophobic patches might inhibit P22 tailspike trimer formation. To test this hypothesis, we measured the ability of 8-anilino-1-naphthalenesulfonic acid (ANS) to inhibit trimerization. ANS is a commonly used probe for measuring exposed hydrophobic residues during formation of both molten globule states and amyloid fibrils (e.g., Souillac et al. 2002; Tanksale et al. 2002). Tailspike protein was refolded in the presence of a final concentration of 1 mM ANS to determine whether ANS is capable of inhibiting trimer formation.

Refolding of denatured tailspike was initiated on ice, and the refolding reaction was incubated for 30 min on ice to allow the initial hydrophobic collapse of the monomer and the formation of the ß-helix domain to occur. Refolding on ice has been shown to improve the yield of refolded trimer (Betts and King 1998) and also helps to populate the protrimer species, aiding in determination of the effects of ANS on all three of the refolding intermediates. Following the 30-min ice incubation, the refolding reaction was treated with 1 mM ANS and samples were collected at various time points, and the reaction was quenched on ice.

The results from this experiment are shown in Figure 7Go. Figure 7AGo shows the outcome of a typical refolding experiment in the absence of ANS. The initial sample is after 30 min on ice. At 0 min, the refolding reaction was shifted from 0° to 20°C. The positions of the monomer, dimer, protrimer, and trimer species are indicated along the edge of the gel. At the initial time point (prior to shifting to 20°C), the sample is composed of predominately dimer with some monomer and protrimer intermediates. The monomer species is relatively short-lived and is no longer detectable after 10 min, whereas the dimer and protrimer both remain detectable for 30 min.



View larger version (53K):
[in this window]
[in a new window]
  		Figure 7. ANS inhibits progression of folding intermediates though the folding pathway. (A) Representative gel demonstrating sample composition ate various time points during a typical refolding reaction. Refolding was initiated on ice, and the reaction was incubated on ice for 30 min prior to shifting to 20°C. Samples were removed at various time points and quenched on ice. The time each sample was removed is indicated below the lane on the gel. Lane 1 (Pre) is the sample prior to shifting to 20°C. Time 0 is immediately following the shift and addition of control buffer. The positions of the three folding intermediates and the trimer on the gel are indicated. (B) Gel demonstrating the effects of addition of 1 mM ANS during refolding. Refolding was initiated on ice and incubated on ice for 30 min. The reaction was shifted to 20°C, and 1 mM ANS was added. Samples were removed at various time points as before. The positions of the three folding intermediates are ndicated along the gel. Note the appearance of shifted intermediate species and aggregate species in the presence of ANS.

 
Figure 7BGo shows a refolding reaction in which the sample was treated with 1 mM ANS at time 0. The most significant effect of the addition of ANS is the inhibition of trimer formation. In the absence of ANS, trimer formation is nearing completion by 240 min. In contrast, there is no detectable trimer after 240 min in the presence of ANS. ANS appears to act by inhibiting the assembly of the various folding intermediates, as indicated by the increased lifetime of both the monomer and dimer intermediates. Monomer is detectable for over 90 min, whereas the dimer species persists for over 4 h, compared to the 10 min and 30 min, respectively seen in the untreated reaction. The protrimer species appears to shift mobility, but appears to exhibit a lifetime similar to that seen in the untreated reaction. However, it appears to convert into either trimer aggregate (which appears in the later timepoints) or higher-order aggregates instead of forming trimer.

All of the intermediates appear to have an altered mobility in the presence of ANS. The monomer and dimer species have a slightly faster mobility, and the protrimer appears to have a slightly retarded mobility. The mobility shift is most likely due to the effects of ANS on the folding intermediates. Congo Red has been shown to perturb tertiary structure and induce aggregation and amyloid fibril formation in immunoglobulin light-chain variable domain dimers (Kim et al. 2003). It is possible that ANS has a similar effect and induces some local structural changes in the tailspike folding intermediates, inducing the altered gel mobilities. An alternative possibility is that binding of ANS to the intermediates results in a change in the surface charge distribution, resulting in the altered mobility without any structural disruption. This effect has been demonstrated with tsf mutations that have altered surface charge (Yu and King 1988). It is currently not possible to distinguish between these two possibilities.

It is clear that the presence of ANS does not cause dissociation of the assembled species and also does not appear to significantly alter the aggregation propensity of the folding intermediates. Either of these possibilities would indicate that the presence of ANS induced large structural perturbations. Instead, monomer and dimer intermediates remain stable in the presence of ANS for extended periods, similar to some cysteine mutants, which have in vitro folding lifetimes of days (Danek and Robinson 2003). Also, a much higher degree of aggregation would be predicted if unfolding of the intermediates was the significant effect of ANS, as aggregation occurs rapidly under unfolded conditions (Lefebvre and Robinson 2003). Because this is not the case, we suspect the primary effect of ANS is inhibition of protein assembly and not induction of a conformational change.

The yield of trimer formation was measured with addition of ANS at different time points during refolding to further characterize the effects of ANS on the formation of trimer. In vitro refolding reactions were performed at 20°C using 14C-labeled full-length P22 tailspike protein. Aliquots from the refolding reaction were treated with varying ANS concentrations (250 µM to 1 mM) at timepoints ranging from 0 to 90 min. The refolding reaction was incubated overnight to reach completion, and the yield of refolded trimer was calculated using nondenaturing gel electrophoresis. A representative gel from this experiment is shown in Figure 8AGo. Figure 8BGo shows the yields of trimer plotted as a function of time when treated with 1 mM ANS. The line at 27 µg/mL shows the average trimer yield in the absence of ANS. In the samples treated with ANS, early-timepoint samples show no trimer formation, whereas samples treated with ANS at later timepoints show increasing yields of refolded trimer. The effectiveness of inhibition of trimer formation is dependent on the time of addition (Fig. 8BGo) and is dose-dependent (Fig. 8CGo), consistent with the hypothesis that burial of hydrophobic residues is critical to trimer formation.



View larger version (29K):
[in this window]
[in a new window]
  		Figure 8. ANS inhibits P22 tailspike trimer formation. (A) Representative gel showing increased wild-type trimer yield with increasingly later addition of ANS to refolding reactions. 1 mM ANS was added to a 100 µg/mL refolding reaction, and the reaction was allowed to reach completion. The time of addition of the ANS is shown along the bottom of the gel. The arrow on the left side of the gel indicates the position of the trimer in the gel. (B) Yields of refolded trimer for the gel shown in panel A are plotted vs. time to demonstrate the decreased effect of ANS over time. The line at 26.8 µg/mL indicates the yield of P22 tailspike trimer in the absence of ANS. Quantitation was performed using Molecular Dynamics’ ImageQuant software package. (C) Representative endpoint trimer yields after ANS addition at 5 min, demonstrating the effect of decreasing ANS concentrations. As can be seen, 1 mM ANS completely inhibits trimerization, whereas 250 µM ANS allows trimerization to occur, though only at 33% of the level of trimerization in the absence of ANS. (D) Denatured tailspike protein was refolded by dilution, and the intrinsic tryptophan fluorescence at 340 nm was monitored over time. The intensity plateaus at ~10 min, indicating that monomer folding has essentially reached completion. The residuals from the first-order data fit are plotted under the graph. (E) Concentration of trimer vs. time ({square}) correlates well with the endpoint yield of trimer following additions of ANS ({blacksquare}). Trimer yields are the average of two experiments. The overlap of the curve indicates that the presence of ANS can affect all stages of the folding pathway.

 
Two methods were used to correlate the decrease of ANS effects with specific steps in the folding pathway. First, to determine whether the effects of ANS on trimerization are due to inhibition of monomer formation, the folding of P22 tailspike monomers was monitored by intrinsic fluorescence. Second, the formation of trimer was determined by nondenaturing gel electrophoresis. Intrinsic tryptophan fluorescence has been commonly used to measure folding of the tailspike monomer species (Fuchs et al. 1991; Miller et al. 1998). When the fluorescence signal was monitored at 340 nm during refolding at 20°C, the signal reached a plateau by 10 min, indicating that monomer folding had reached completion (Fig. 8DGo). As Figure 8BGo shows, ANS inhibits trimer formation well beyond 10 min. This indicates that the effect of ANS on trimer formation was not purely due to inhibition of the monomer folding by the presence of ANS, though this surely does play a role in the complete inhibition seen at early timepoints.

Fluorescence measurements provide a measure of the folding of the monomer, but they provide no information about the later folding and assembly events. The presence of various folding intermediates over time was determined using nondenaturing gel electrophoresis and quantitated using radiolabeled P22 tailspike to correlate the effects of ANS with the folding pathway. The amount of trimer present at each timepoint in the refolding reaction was proportional to the amount of trimer formed in the presence of 1 mM ANS at each timepoint (Fig. 8EGo). These results, combined with the results in Figure 7Go, indicate that ANS acts by blocking both the assembly of the protrimer and the transition of protrimer to trimer.


   Discussion
TOP
 Abstract
 Introduction
 Results
 Discussion
Materials and methods
 References
 
Folding and assembly of multimeric proteins is a complicated process. In some cases, folding occurs first and then assembly occurs, as is often the case for viruses and enzymes (Tellinghuisen et al. 2001; Zhou and Weiner 2001). In other cases, the final folded polypeptide requires certain assembly events to occur during folding, either simultaneously or at certain checkpoints along the folding pathway. Examples of coordinated folding and assembly include, but are not restricted to, coiled-coil domains (Burkhard et al. 2001), bacterial luciferase (Baldwin et al. 1993), and the P22 tailspike (Fuchs et al. 1991). Decoupling of the folding and assembly processes is very difficult in proteins that have coordinated folding and assembly processes.

C terminus of the P22 tailspike acts as oligomerization domain
Study of the P22 tailspike protein has provided useful insight into coordinated folding and assembly processes. Nondenaturing polyacrylamide gel electrophoresis has successfully been used to identify intermediates along the folding and aggregation pathways (Fig. 2Go). In the current model for folding and assembly of P22 tailspike, unfolded polypeptides fold into assembly-competent monomer species that are sequentially added together to form the protrimer species, which then undergoes structural rearrangements to form the final folded product. It has been proposed that the C terminus is critical for assembly, due to the large hydrophobic patches in that region.

Miller et al. (1998) reported that the isolated ß-helix domain forms a concentration-dependent equilibrium between monomer and trimer, implying that there are weak interactions within the helix domain that might play a role in trimer formation. However, their results indicated that protrimer formation required more than association along the axis of the ß-helix. In addition, Kreisberg et al. (2002) demonstrated that the G546N mutation stalls folding as a protrimer species, whereas R563Q and A575T stall as monomers. These results support the hypothesis that the C terminus is critical for formation of the critical protrimer intermediate (Kreisberg et al. 2002).

The present findings further show that the C terminus plays a critical role in protrimer formation, and that the hydrophobic residues within the C terminus appear to provide a driving force for assembly. First, both the TSP{Delta}C and TSP{Delta}NC truncations exhibit weak monomer-trimer equilibrium, implying that there is only a weak interaction between the ß-helix domains. Second, the full-length tailspike protein formed mixed species with the TSP{Delta}N truncation. However, the TSP{Delta}C and TSP{Delta}NC truncations were unable to form mixed species with the full-length protein. Third, mutations to the hydrophobic core of the C terminus block trimer formation. Finally, high concentrations of ANS inhibit trimer formation by inhibiting assembly of folding intermediates and by blocking the protrimer-to-trimer transition. The effects of ANS correlate with the level of trimer present in the refolding reaction. It was previously shown that GroEL inhibits trimer formation at 35°C, in a manner similar to that reported here for ANS (Brunschier et al. 1993). In addition, the majority of lethal mutations are found in the C terminus of the protein (Beibinger et al. 1995). These results provide a significant amount of evidence to support the hypothesis that the C terminus and specifically the hydrophobic regions within the C terminus act as the nucleation point for trimer assembly.

Assembly of the protrimer
The results of this work provide additional insights into the folding pathway presented in Figure 2Go. The initial step in refolding is formation of the partially structured monomer (U->M). This transition is believed to consist of the formation of the ß-helix domain (Danner and Seckler 1993; Miller et al. 1998; Schuler and Seckler 1998), which can be monitored by both tryptophan fluorescence and CD (Danner and Seckler 1993). Similar measurements of the conformation of the N and C termini in this partially structured monomer have not been reported. The N and C termini are typically represented as disordered in the early stages of refolding in most schematics of the folding pathway of P22 tailspike. The present results show that the C terminus provides a significant driving force for trimerization and likely folds at an early point in the folding pathway.

Figure 9Go shows a revised schematic for P22 tailspike refolding in which the C terminus is depicted as an ordered species. The ß-sheets D and E in each monomer form the ß-prism and caudal fin domains respectively, and the data we present here suggest that they may form during the early stages in refolding, while the ß-helix domain is also forming. The D and E sheets of one monomer likely interact with each other to shield their hydrophobic faces from water. The weak interactions between the ß-helix domains may initiate the initial association between monomers, which is then stabilized by rearrangement of the D and E sheets to interact with the complementary sheet on the second subunit, though in nonnative interactions.



View larger version (18K):
[in this window]
[in a new window]
  		Figure 9. Revised in vitro refolding pathway. Unfolded monomer folds and forms either aggregate-prone monomer or folding-competent monomer. The two species may be in equilibrium, or environmental factors may induce formation of one species over the other. Two monomers are sequentially added together to form either a folding-competent dimer or dimer aggregate. In the case of a folding-competent dimer, the associations are formed through the C terminus, as shown. Addition of a third monomer through C-terminal interactions leads to formation of the protrimer species, which then forms the fully folded trimer.

 
Nonnative hydrophobic interactions play an important role in the folding of the colicin immunity protein Im7. Im7 is a four-helix protein that rapidly forms a stable, structured intermediate during in vitro folding (Capaldi et al. 2002). Helices I, II, and IV form first and then interact in a nonnative manner to shield their hydrophobic faces until helix III can be formed, leading to the native conformation (Capaldi et al. 2002). An analogous mechanism may be occurring during assembly of the P22 tailspike trimer, where nonnative interactions have been shown to exist in the protrimer (Robinson and King 1997; Benton et al. 2002). The C terminus most likely forms a molten globule state in the dimer and protrimer intermediates where ß-sheets D and E are formed but in which the contacts between the sheets are much looser than in the native structure.

Assembly through hydrophobic interactions has a number of advantages. One advantage is that burying of hydrophobic residues provides a driving force for assembly. A second advantage is that hydrophobic interactions, unlike ionic interactions, are able to shift with little energetic cost. In the case of tailspike, this could allow the structural rearrangements during the protrimer-to-trimer transition to occur without the energetic cost associated with disrupting ionic interactions. The mobility offered by hydrophobic residues, however, does not ensure that the subunits end up in proper register with each other. This may be the function of the transient disulfide bond that occurs during folding (Robinson and King 1997; Haase-Pettingell et al. 2000). The transient disulfide bond has been shown to exist in the dimer species (B. Danek and A. Robinson, unpubl.), though it is currently unclear whether the disulfide forms during dimer formation or after the dimer has already formed.

This model also provides some additional insight into the mutational results reported by Kreisberg et al. (2002). The G546D mutation is in the intertwined region of the trimer, and the R563Q and A575T mutations are both on the D ß-sheet, which forms the ß-prism. The R563Q mutation disrupts an ionic interaction between R563 and D572, potentially disrupting the formation of the D ß-sheet. The substitution of threonine at position 575 inserts a bulkier side chain into the ß-strand, which may interrupt critical hydrogen bond interactions along that region of the ß-sheet, also disrupting the formation of the D ß-sheet. In contrast, the G546D mutation is not located on any of the strands that form the D ß-sheet and therefore would not disrupt this sheet. This is one possible reason why G546D is capable of forming a stalled protrimer and the other species are unable to do so.

One interesting result is that the R563Q and A575T mutations form soluble monomer species, whereas the mutations described in the present report all form insoluble monomers. Kreisberg et al. (2002) proposed that the C terminus might interact with the ß-helix in the R563Q and A575T mutants to shield the hydrophobic residues that are presumably exposed, creating soluble monomeric species. An alternative hypothesis is that the D and E ß-sheets fold onto each other after formation to shield the hydrophobic regions. This would explain why the mutations reported by Kreisberg et al. (2002) remain soluble and those detailed here are insoluble. The side chains for the three mutants studied by Kreisberg and colleagues are all on the exterior face of the ß-prism sheets, which would not disrupt hydrophobic interactions between the caudal fin domain and the ß-prism domain. In contrast, the side chains for the mutations we studied here point into the hydrophobic core, which would disrupt any intrachain hydrophobic interactions as well as interchain hydrophobic interactions, causing prolonged exposure of hydrophobic residues, promoting aggregation. This intramolecular hydrophobic interaction may occur as part of the folding pathway as well, preventing aggregation until intermolecular interactions can form.

These results also provide some insight into the in vivo folding of P22 tailspike. The full-length nascent polypeptide chain has been shown to remain in contact with the ribosome for considerable lengths of time (Clark and King 2001). It is possible that the ribosome may act as a chaperone, shielding the hydrophobic residues until a second polypeptide chain can be found and a dimer can be formed.

Implications for general protein assembly
The mechanism proposed is one that is not without precedence in nature. The folding pathway of collagen occurs through a mechanism similar to that proposed here. The basic building block of collagen is the procollagen molecule, which is similar to the folded monomer in the P22 tailspike folding pathway. The procollagen is synthesized and assembles to form the collagen trimer through a C to N terminus zipper-like process (Bachinger et al. 1980), like the one proposed here for P22 tailspike. Proper alignment of the three subunits is maintained through the formation of intersubunit disulfide bonds that form prior to formation of the collagen triple helix (Bachinger et al. 1980), in a mechanism similar to the one we propose here.

Protein–protein interfaces in oligomeric proteins and in multiprotein complexes are commonly stabilized by salt bridges and hydrogen bonds (Miller 1989; Janin and Chothia 1990), and the cores of the individual subunits are often hydrophobic. The P22 tailspike protein is an example where both types of interactions play a role in the folding of the protein. The interface between the ß-helix domains is dominated by hydrogen bonding interactions between water and the adjacent subunits, and each helix is stabilized by the predominantly hydrophobic interactions in the center of the helix. In contrast, the interface between the three chains in the C terminus is dominated by hydrophobic interactions and resembles the core of a protein more than a typical protein–protein interface. The data presented here suggest that hydrophobic interfaces may play an important role in the assembly of some multimeric proteins.


   Materials and methods
TOP
 Abstract
 Introduction
 Results
 Discussion
 Materials and methods
References
 
Materials
All chemicals used were obtained from major commercial suppliers. The 14C-labeled L-amino acid mixture was from Perkins Elmer. Restriction enzymes were from New England Biolabs. Primers used for cloning and mutagenesis were from both Genelink and Integrated DNA Technologies.

Site-directed mutagenesis
Mutagenesis was performed using the QuikChange Site-Directed Mutagenesis kit (Stratagene). Gene-specific primers containing either the specific mutations or degenerate bases to generate a series of mutations were designed. Mutagenesis was performed according to the manufacturer’s instructions using a pET11a vector containing the full-length P22 tailspike gene as a template. The identity of each mutation was verified by complete sequencing of the gene.

Construction of tailspike truncations
Gene-specific primers containing an NdeI restriction site 5' to the first amino acid were designed to anneal to the N terminus and at residue 109. Reverse primers containing an EcoRI site 3' to the terminal amino acids were generated to anneal at the C terminus and at residue 544. Truncated genes were generated by PCR using a pET11a vector containing the full-length P22 tailspike gene using an MTC-200 Peltier Thermal Cycler (MJ Research). Truncated tailspike genes and pET11a vector were digested by NdeI and EcoRI restriction enzymes. Digested genes and the pET11a backbone were ligated overnight at 16°C using T4 ligase prior to transformation into chemically competent E. coli DH5{alpha} cells. Each truncation was verified by complete sequencing.

Protein expression
Chemically competent E. coli BL21(DE3) cells (Novagen) were transformed with a pET11a plasmid containing the appropriate gene and selected for on LB-Ampicillin plates. Individual colonies were grown in LB media (Sambrook et al. 1989) to an OD600 ~0.5 at 30°C in LB media and induced using 1 mM IPTG for 4 to 24 h. Radioactive protein was produced as described (Danek and Robinson 2003). Cells were harvested by centrifugation and resuspended in lysis buffer (50 mM Tris, 5 mM MgCl2, 0.1% Triton X-100, 0.1 mg/mL lysozyme, 0.1 mg/mL DNase). After two freeze/thaw cycles (-80°C/20°C), the cell debris was removed by centrifugation at 12,000g. The supernatant and resulting pellet were separated for SDS-PAGE analysis. The pellet was resuspended in a volume of lysis buffer equivalent to the supernatant prior to analysis.

Protein purification
Large-scale protein preparation of full-length or truncated P22 tailspike or 14C-labeled tailspike was conducted as described (Danek and Robinson 2003; Lefebvre and Robinson 2003). Fractions from the hydroxyapatite column containing tailspike were collected and concentrated to 10–15 mg/mL using Vivaspin 20-mL concentrators (Vivascience). Purity of > 95% was determined by SDS-PAGE for each of the proteins. Final concentration was determined by measuring the OD278 (1 OD278 = 1.017 mg/mL tailspike, 1.11 mg/mL TSP{Delta}N, 1.19 mg/mL TSP{Delta}C, 1.33 mg/mL TSP{Delta}NC).

Circular dichroism
Purified TSP{Delta}C was diluted in 0.01X Tris Refolding Buffer (50 mM Tris, pH 7.6, 2 mM EDTA) to 100 µg/mL and placed in a quartz cuvette with a optical path length of 1 cm. The spectrum was measured from 260 to 190 nm using an Aviv Model 215 Circular Dichroism Spectrometer using 2 nm steps and a 5-sec integration time.

Analytical ultracentrifugation
Purified TSP{Delta}C was diluted in 50 mM Tris, pH 7.6, 25 mM NaCl to 2 µM, 6 µM, and 8 µM. Sedimentation equilibrium runs were performed using a Beckman Optima XL-A Analytical Ultracentrifuge at 5°C and 10,000, 12,500, and 15,000 rpm. Samples were equilibrated for 999 min at each speed and then scanned two times with 120 min between scans to ensure that the samples had reached equilibrium. Data were analyzed using the program NonLin, which is available through the RASMB website (http://www.bbri.org/RASMB/rasmb.html). A partial specific volume of 0.7337 for the protein was calculated using the program Sednterp (also available through the RASMB website).

Fluorescence measurements
Purified TSP{Delta}C was diluted in 1X Tris Refolding Buffer to 100 µg/mL. Fluorescence spectra were measured from 300 nm to 400 nm (excitation wavelength, 280 nm) using a Hitachi F-4500 Fluorescence Spectrophotometer. Center of mass was calculated using the formula:


where I is equal to intensity and |gn is equal to the wavenumber. Fluorescent kinetics were measured by rapidly mixing 25 µL of denatured tailspike with 2475 µL of Tris Refolding Buffer in a quartz cuvette treated with 5% Tween. Purified P22 tailspike protein was denatured at 1 mg/mL in 8 M Tris-acid urea for 1 h at room temperature. The sample was excited at 280 nm, and fluorescence was monitored at 340 nm for ~1 h. Data were fit as described (Danek and Robinson 2003).

Tailspike refolding
Tailspike or truncated tailspike proteins were denatured in 8 M Tris-Acid urea, pH 3.0, for 1 h at either 0°C or 20°C at a concentration of 1 mg/mL. Denaturation temperature was identical to initial refolding temperature. Refolding was initiated by dilution of the denatured protein into Tris Refolding Buffer to the final experimental concentration in buffer pre-incubated at the initial temperature. Samples in which refolding was initiated at ~0°C were incubated on ice for 30 min and then placed at the final refolding temperature. Samples for endpoint determination were incubated overnight at 20°C. Samples for the timepoint experiment were incubated at 20°C, and samples were removed at various timepoints and quenched on ice in 3X nondenaturing sample buffer (15 mM Tris, 120 mM glycine, 30% glycerol, bromophenol blue). Previous studies (Speed et al. 1995) showed that no changes in the intermediates occur during this incubation. Data were fit to a first-order rate as described (Danek and Robinson 2003).

ANS inhibition
For timepoint experiments, refolding was initiated at 0°C as described above. After 30 min on ice, the sample was treated with 10 mM ANS to a final concentration of 1 mM or an equivalent amount of control buffer, and aliquots are removed at various timepoints. For endpoint experiments, refolding was initiated at 20°C as described above. At specific timepoints, samples were mixed with varying concentrations of ANS, and refolding was allowed to continue in the presence of ANS overnight.

Electrophoresis and staining
Nondenaturing polyacrylamide gel electrophoresis was performed at 4°C as described (Speed et al. 1995). Resolving and stacking gels were 7% and 9% for ANS inhibition and mixed refolding, respectively. Silver staining was performed as described (Sather and King 1994). Molecular Dynamics Phosphorimager plates were exposed to dried gels containing radiolabeled protein for 3–5 d, and then scanned using a Molecular Dynamics Phosphorimager plate reader and quantitated using ImageQuant software.


   Acknowledgments
 
We thank Junghwa Kim for the generous contribution of 14C-labeled tailspike protein; Eugene Antipov for assistance in generating the suppressor double mutants; Dr. Joel Schneider, Juliana Kretsinger, Dr. Karen Fleming, and Lumelle Schneeweis for assistance collecting and analyzing the ultracentrifugation data; and Brenda Danek, Junghwa Kim, and Dr. Yong Duan for helpful discussions during preparation of this manuscript. This work was supported in part by NIH grant P20 RR15588.

The publication costs of this article were defrayed in part by payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 USC section 1734 solely to indicate this fact.


   References
TOP
 Abstract
 Introduction
 Results
 Discussion
 Materials and methods
 References
 
Bachinger, H.P., Bruckner, P., Timpl, R., Prockop, D.J., and Engel, J. 1980. Folding mechanism of the triple helix in type-III collagen and type-III pN-collagen. Role of disulfide bridges and peptide bond isomerization. Eur. J. Biochem. 106: 619–632.[Medline]

Baldwin, T.O., Ziegler, M.M., Chaffotte, A.F., and Goldberg, M.E. 1993. Contribution of folding steps involving the individual subunits of bacterial luciferase to the assembly of the active heterodimeric enzyme. J. Biol. Chem. 268: 10766–10772.[Abstract/Free Full Text]

Ban, N., Nissen, P., Hansen, J., Moore, P.B., and Steitz, T.A. 2000. The complete atomic structure of the large ribosomal subunit at 2.4 Å resolution. Science 289: 905–920.[Abstract/Free Full Text]

Beibinger, M., Lee, S.C., Steinbacher, S., Reinemer, P., Huber, R., Yu, M.-H., and Seckler, R. 1995. Mutations that stabilize folding intermediates of phage P22 tailspike protein: Folding in vivo and in vitro, stability, and structural context. J. Mol. Biol. 249: 185–194.[CrossRef][Medline]

Benton, C.B., King, J., and Clark, P.L. 2002. Characterization of the protrimer intermediate in the folding pathway of the interdigitated ß-helix tailspike protein. Biochemistry 41: 5093–5103.[CrossRef][Medline]

Betts, S.D. and King, J. 1998. Cold rescue of the thermolabile tailspike intermediate at the junction between productive folding and off-pathway aggregation. Protein Sci. 7: 1516–1523.[Abstract]

Brunschier, R., Danner, M., and Seckler, R. 1993. Interactions of phage P22 tailspike protein with GroE molecular chaperones during refolding in vitro. J. Biol. Chem. 268: 2767–2772.[Abstract/Free Full Text]

Buevich, A. and Baum, J. 2001. Nuclear magnetic resonance characterization of peptide models of collagen-folding diseases. Philos. Trans. R. Soc. Lond. B. Biol. Sci. 356: 159–168.[CrossRef][Medline]

Buratowski, R.M., Downs, J., and Buratowski, S. 2002. Interdependent interactions between TFIIB, TATA binding protein, and DNA. Mol. Cell. Biol. 22: 8735–8743.[Abstract/Free Full Text]

Burkhard, P., Stetefeld, J., and Strelkov, S.V. 2001. Coiled coils: A highly versatile protein folding motif. Trends Cell Biol. 11: 82–88.[CrossRef][Medline]

Burkhard, P., Ivaninskii, S., and Lustig, A. 2002. Improving coiled-coil stability by optimizing ionic interactions. J. Mol. Biol. 318: 901–910.[CrossRef][Medline]

Capaldi, A.P., Kleanthous, C., and Radford, S.E. 2002. Im7 folding mechanism: Misfolding on a path to the native state. Nat. Struct. Biol. 9: 209–216.[Medline]

Clark, P.L. and King, J. 2001. A newly synthesized, ribosome-bound polypeptide chain adopts conformation dissimilar from early in vitro refolding intermediates. J. Biol. Chem. 276: 25411–25420.[Abstract/Free Full Text]

Danek, B.L. and Robinson, A.S. 2003. Non-native interactions between cysteines direct productive assembly of P22 tailspike protein. Biophys. J. 85: (in press)

Danner, M. and Seckler, R. 1993. Mechanism of phage P22 tailspike folding mutations. Protein Sci. 2: 1869–1881.[Abstract]

Danner, M., Fuchs, A., Miller, S., and Seckler, R. 1993. Folding and assembly of phage P22 tailspike protein lacking N-terminal, head-binding domain. Eur. J. Biochem. 215: 653–661.[Medline]

Fuchs, A., Seiderer, C., and Seckler, R. 1991. In vitro folding pathway of phage P22 tailspike protein. Biochemistry 30: 6598–6604.[CrossRef]