Abstract: The GroEL/GroES chaperonin system mediates protein folding in the bacterial cytosol. Newly synthesized proteins reach GroEL via transfer from upstream chaperones such as DnaK/DnaJ (Hsp70). Here we employed single molecule and ensemble FRET to monitor the conformational transitions of a model substrate as it proceeds along this chaperone pathway. We find that DnaK/DnaJ stabilizes the protein in collapsed states that fold exceedingly slowly. Transfer to GroEL results in unfolding, with a fraction of molecules reaching locally highly expanded conformations. ATP-induced domain movements in GroEL cause transient further unfolding and rapid mobilization of protein segments with moderate hydrophobicity, allowing partial compaction on the GroEL surface. The more hydrophobic regions are released upon subsequent protein encapsulation in the central GroEL cavity by GroES, completing compaction and allowing rapid folding. Segmental chain release and compaction may be important in avoiding misfolding by proteins that fail to fold efficiently through spontaneous hydrophobic collapse.

Abstract: The GroEL/GroES chaperonin system of Escherichia coli forms a nano-cage allowing single protein molecules to fold in isolation. However, as the chaperonin can also mediate folding independently of substrate encapsulation, it remained unclear whether the folding cage is essential in vivo. To address this question, we replaced wild-type GroEL with mutants of GroEL having either a reduced cage volume or altered charge properties of the cage wall. A stepwise reduction in cage size resulted in a gradual loss of cell viability, although the mutants bound non-native protein efficiently. Strikingly, a mild reduction in cage size increased the yield and the apparent rate of green fluorescent protein folding, consistent with the view that an effect of steric confinement can accelerate folding. As shown in vitro, the observed acceleration of folding was dependent on protein encapsulation by GroES but independent of GroES cycling regulated by the GroEL ATPase. Altering the net-negative charge of the GroEL cage wall also strongly affected chaperonin function. Based on these findings, the GroEL/GroES compartment is essential for protein folding in vivo.

Abstract: gp120 is a subunit of the envelope glycoprotein of HIV-1. The third variable loop region of gp120 (V3 loop) contains multiple immunodominant epitopes and is also functionally important for deciding cell-tropism of the virus. 447-52D is a monoclonal antibody that recognizes the conserved tip of the V3 loop in a beta-turn conformation. This antibody has previously been shown to neutralize diverse strains of the virus. In an attempt to generate an immunogen competent to generate 447-52D-like antibodies, the known epitope of 447-52D was inserted at three different surface loop locations in the small, stable protein Escherichia coli Trx (thioredoxin). At one of the three locations (between residues 74 and 75), the insertion was tolerated, the resulting protein was stable and soluble, and bound 447-52D with an affinity similar to that of intact gp120. Upon immunization, the V3 peptide-inserted Trx scaffold was able to generate anti-V3 antibodies that could compete out 447-52D binding to gp120. Epitope mapping studies demonstrated that these anti-V3 antibodies recognized the same epitope as 447-52D. Although the 447-52D-type antibodies were estimated to be present at concentrations of 50-400 microg/ml of serum, these were not able to effect neutralization of strains like JRFL and BAL but could neutralize the sensitive MN strain. The data suggest that because of the low accessibility of the V3 loop on primary isolates such as JRFL, it will be difficult to elicit a V3-specific, 447-52D-like antibody response to effectively neutralize such isolates.

Abstract: CD4 is an important component of the immune system and is also the cellular receptor for HIV-1. CD4 consists of a cytoplasmic tail, one transmembrane region, and four extracellular domains, D1-D4. Constructs consisting of all four extracellular domains of human CD4 as well as the first two domains (CD4D12) have previously been expressed and characterized. All of the gp120-binding residues are located within the first N-terminal domain (D1) of CD4. To date, it has not been possible to obtain domain D1 alone in a soluble and active form. Most residues in CD4 that interact with gp120 lie within the region 21-64 of domain D1 of CD4. On the basis of these observations and analysis of the crystal structure of CD4D12, a mutational strategy was designed to express CD4D1 and region 21-64 of CD4 (CD4PEP1) in Escherichia coli. K(D) values for the binding of CD4 analogues described above to gp120 were measured using a Biacore-based solution-phase competition binding assay. Measured K(D) values were 15 nM, 40 nM, and 26 microM for CD4D12, CD4D1, and CD4PEP1, respectively. All of the proteins interact with gp120 and are able to expose the 17b-binding epitope of gp120. Structural content was determined using CD and proteolysis. Both CD4D1 and CD4PEP1 were partially structured and showed an enhanced structure in the presence of the osmolyte sarcosine. The aggregation behavior of all of the proteins was characterized. While CD4D1 and CD4PEP1 did not aggregate, CD4D12 formed amyloid fibrils at neutral pH within a week at 278 K. These CD4 derivatives should be useful tools in HIV vaccine design and entry inhibition studies.

Abstract: Single-chain derivatives of JRFL gp120 linked to the first two domains of human CD4 (gp120-CD4D12) or to the CD4 miniprotein analog CD4M9 (gp120-M9), have been constructed. Biacore studies revealed that gp120-CD4D12 and gp120-M9 bound to antibody 17b with dissociation constants of 0.8 and 25 nM, respectively, at pH 7.0, while gp120 alone did not bind. The binding of gp120-CD4D12 to 17b is not affected by the addition of excess soluble CD4D12, while the binding of gp120-M9 is enhanced. This finding indicates that the M9 component of the single chain interacts relatively weakly with gp120 and can be displaced by soluble CD4D12. Immunogenicity studies of gp120, gp120-CD4D12, and gp120-M9 were carried out with guinea pigs. All three molecules were highly immunogenic. The resulting antisera were examined for neutralizing activities against various human immunodeficiency virus type 1 isolates. Broadly neutralizing activity was observed only with sera generated against gp120-CD4D12. These antisera were depleted of anti-CD4D12 antibodies by being passed over a column containing immobilized CD4D12. The depleted sera showed a loss of broadly neutralizing activity. Sera that were affinity purified over a column containing immobilized gp120-M9 also lacked such neutralizing activity. This finding suggests that the broadly neutralizing response observed is exclusively due to anti-CD4 antibodies. Competition experiments showed that only antisera generated against gp120-CD4D12 competed with the CD4i antibody 17b and that this activity was not affected by depletion of anti-CD4 antibodies. The data indicate that although antibodies targeting the CD4i epitope were generated by the gp120-CD4D12 immunogen, these antibodies were nonneutralizing.

Abstract: Disulfides cross-link residues in a protein that are separated in primary sequence and stabilize the protein through entropic destabilization of the unfolded state. While the removal of naturally occurring disulfides leads to protein destabilization, introduction of engineered disulfides does not always lead to significant stabilization of a protein. We have analyzed naturally occurring disulfides that span adjacent antiparallel strands of beta sheets (cross-strand disulfides). Cross-strand disulfides have recently been implicated as redox-based conformational switches in proteins such as gp120 and CD4. The propensity of these disulfides to act as conformational switches was postulated on the basis of the hypothesis that this class of disulfide is conformationally strained. In the present analysis, there was no evidence to suggest that cross-strand disulfides are more strained compared to other disulfides as assessed by their torsional energy. It was also observed that these disulfides occur solely at non-hydrogen-bonded (NHB) registered pairs of adjacent antiparallel strands and not at hydrogen-bonded (HB) positions as suggested previously. One of the half-cystines involved in cross-strand disulfide formation often occurs at an edge strand. Experimental confirmation of the stabilizing effects of such disulfides was carried out in Escherichia coli thioredoxin. Four pairs of cross-strand cysteines were introduced, two at HB and two at NHB pairs. Disulfides were formed in all four cases. However, as predicted from our analysis, disulfides at NHB positions resulted in an increase in melting temperature of 7-10 degrees C, while at HB positions there was a corresponding decrease of -7 degrees C. The reduced state of all proteins had similar stability.

Abstract: gp120 is a subunit of the Env (viral envelope protein) of HIV-1. The protein consists of inner and outer domains linked by a bridging sheet. Several gp120 residues that bind the neutralizing antibody 17b as well as the cellular co-receptor CCR5 (CC chemokine receptor 5), are located in the bridging sheet. Peptides that mimic the 17b-binding regions of gp120 would be useful potential immunogens for the generation of neutralizing antibodies against HIV-1. Towards this end, a 26-residue, four-stranded beta-sheet peptide was designed on the basis of the structure of the bridging sheet, and its structure was characterized in methanol by NMR. In methanol, amide and alpha-proton resonances were well resolved and dispersed. A number of interstrand NOEs (nuclear Overhauser effects) were observed, providing good evidence for multiple turn beta-hairpin structure. NOEs also provided good evidence for all Xxx-D-Pro bonds in the trans configuration and all three turns formed by a two residue D-Pro-Gly segment to be of type II' turn. The structure conforms well to the designed four-stranded beta-sheet structure. Approx. 20% of the peptide was estimated to adopt a folded conformation in water, as evidenced by CD spectroscopy. This was consistent with smaller, but still significant, downfield shifts of C(alpha)H protons relative to random-coil values. A second peptide was designed with two disulphide bonds to further constrain the peptide backbone. While structured in methanol, this peptide, like the previous one, also exhibits only partial structure formation in water, as evidenced by CD spectroscopy.

Abstract: Accurate prediction of location of cavities and surface grooves in proteins is important, as these are potential sites for ligand binding. Several currently available programs for cavity detection are unable to detect cavities near the surface or surface grooves. In the present study, an optimized molecular dynamics based procedure is described for detection and quantification of interior cavities as well as surface pockets. This is based on the observation that the mobility of water in such pockets is significantly lower than that of bulk water. The algorithm efficiently detects surface grooves that are sites of protein-ligand and protein-protein interaction. The algorithm was also used to substantially improve the performance of an automated docking procedure for docking monomers of nonobligate protein-protein complexes. In addition, it was applied to predict key residues involved in the binding of the E. coli toxin CcdB with its inhibitor. Predictions were subsequently validated by mutagenesis experiments.