Retinoblastoma Protein Binding Properties Are Dependent on 4 Cysteine Residues in the Protein Binding Pocket*

The retinoblastoma gene product (pRB) participates in regulating mammalian cell replication. The mechanism responsible for pRB’s growth regulatory activity is uncertain. However, pRB is known to bind viral transforming proteins including the papilloma virus E7 protein, cellular proteins, and DNA. pRB contains a critical domain termed the “binding pocket” which is required for binding activities. This binding pocket contains 8 cysteine residues. A naturally occurring mutation affecting one of these cysteines is known to eliminate pRB’s protein and DNA binding activities. To investigate the cysteine residues in pRB’s binding pocket, each residue was mutated to alanine, phenyl- alanine, or serine. These mutant genes were used to prepare pRBs harboring specific amino acid substitu- tions. Individual mutations at positions 407, 553,666, and 706 depressed pRB binding to E7 protein, DNA, and a conformation-specific anti-pRB antibody, XZ133. Combinations of these inhibitory mutations exhibited additive inhibitory effects on pRB’s binding properties. Mutations at positions 438, 489, 590, 712, and 853 did not affect pRB binding to E7 protein, DNA, or the XZ133 antibody. Combination of these five neutral mutations yielded a pRB species with full E7 pro- tein, DNA, and XZ 133 binding activities. These studies indicate that the cysteine residues at positions 407, 553, 666, and 706 contribute to the E7 protein and DNA binding properties of pRB and appear to do so by maintaining pRB’s

The retinoblastoma gene product (pRB) participates in regulating mammalian cell replication. The mechanism responsible for pRB's growth regulatory activity is uncertain. However, pRB is known to bind viral transforming proteins including the papilloma virus E7 protein, cellular proteins, and DNA. pRB contains a critical domain termed the "binding pocket" which is required for binding activities. This binding pocket contains 8 cysteine residues. A naturally occurring mutation affecting one of these cysteines is known to eliminate pRB's protein and DNA binding activities. To investigate the cysteine residues in pRB's binding pocket, each residue was mutated to alanine, phenylalanine, or serine. These mutant genes were used to prepare pRBs harboring specific amino acid substitutions. Individual mutations at positions 407, 553,666, and 706 depressed pRB binding to E7 protein, DNA, and a conformation-specific anti-pRB antibody, XZ133. Combinations of these inhibitory mutations exhibited additive inhibitory effects on pRB's binding properties. Mutations at positions 438, 489, 590, 712, and 853 did not affect pRB binding to E7 protein, DNA, or the XZ133 antibody. Combination of these five neutral mutations yielded a pRB species with full E7 protein, DNA, and XZ 133 binding activities. These studies indicate that the cysteine residues at positions 407, 553, 666, and 706 contribute to the E7 protein and DNA binding properties of pRB and appear to do so by maintaining pRB's normal conformation.
The retinoblastoma suppressor gene product is a 105-kDa phosphoprotein (pRB105) (1,2) which is thought to play a key role in regulating mammalian cell replication (3)(4)(5). The wild type retinoblastoma (RB)' gene is normally expressed in all body tissues. However, mutated or partially deleted versions of the RB gene have been found in a variety of human tumors and tumor cell lines (6)(7)(8)(9). Introduction of normal RB genes into these tumor cell lines retards their growth and inhibits their ability to form tumors in nude mice (10,11). The association of tumors with the loss of normal RB alleles has given rise to t h e hypothesis that RB's normal function is to inhibit cell proliferation. The precise mechanism employed by the RB gene product (pRB) to block cell replication is * The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "aduertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
The abbreviations used are: RB, retinoblastoma protein; ELISA, enzyme-linked immunosorbent assay; HEPES, 4-(2-hydroxyethyl)-lpiperazineethanesulfonic acid; SDS, sodium dodecyl sulfate; PAGE, polyacrylamide gel electrophoresis; HPV, human papilloma virus. unknown. However, two general biochemical properties of RB proteins have been described. First, pRB can form specific complexes with a variety of proteins including the viral transforming proteins of SV40, adenovirus, and human papilloma virus (HPV): large-T, ElA, and E7, respectively (12)(13)(14). pRB also binds normal cellular proteins (15)(16)(17) and has recently been shown to associate with the E2F/DRTF transcription factor complex (18)(19)(20). The segment of pRB that binds to these proteins consists of two discontinuous regions designated domains A and B, which encompass amino acids 394-571 and 649-772, respectively (21, 22). These regions have been proposed to form a "binding pocket" which is essential for binding of viral oncoproteins and normal cellular proteins to pRB. The second biochemical property associated with pRB is its ability to bind double-stranded DNA. RB protein binds to DNA in a non-sequence-specific manner which is commonly demonstrated using random fragments of calf thymus DNA (23-25). The functional significance of pRB's DNA binding activity is unknown.
Both full-length pRB105 and a smaller 60-kDa version of the RB protein (pRB6O) contain the protein binding pocket and exhibit DNA binding activity. The critical amino acid residues within these pRB species that govern their biochemical properties are as yet undefined. Nonetheless, two observations call attention to the cysteine residues within the RB protein binding pocket. First, a naturally occurring mutant form of pRB has been found which contains a point mutation that converts Cys-706 to phenylalanine. This mutant protein fails to bind viral transforming proteins and lacks cell growthinhibitory activity in vitro (26,27). Second, examination of the amino acid sequence of the related adenovirus E1A and HPV-16 E7-binding protein, p107, suggests that these residues are important (28). Four of the 8 cysteine residues within the pRB6O binding pocket are conserved in p107. The biochemical mechanism underlying the importance of these residues is unclear. However, cysteines commonly contribute to structural interactions via hydrogen bonds, disulfide bonds, or coordinate metal ion binding. The presence of critical cysteine residues within the binding pocket of pRB suggests that one or more of these biochemical activities may be important for maintaining the functional integrity of RB proteins.
To characterize the contribution of each cysteine residue in the pRB binding pocket to the biochemical function of RB protein, individual cysteine residues were mutated to alanine or other amino acids. A series of altered RB proteins were obtained from these mutated genes and analyzed for their abilities to bind HPV-16 E7 protein, DNA, and a conformation-specific anti-pRB antibody.

MATERIALS AND METHODS
Plasmids-The DNA sequence encoding human pRB was cloned from a human lung fibroblest library (Clontech, Palo Alto, CA) using oligonucleotide probes derived from a published RB nucleic acid sequence (1,29). The cloning of the 60-kDa fragment of pRB105 which begins at Met-387 has been described previously (30). The substitution mutations were made by site-directed mutagenesis using oligonucleotides and the pSelect mutagenesis system (Promega, Madison, WI). Each of the substitution mutations were placed in pGem4Z (Promega) for in vitro transcription-translation. Combinations of substitution mutants were made by either successive site-directed mutagenesis or by swapping restriction fragments containing the desired single mutation. All mutations were confirmed by DNA sequencing.
Recombinant HPV-16 E7"Tbe HPV-16 E7 gene (P. M. Howley, National Cancer Institute, Bethesda, MD) was cloned in a Tac promoter bacterial expression plasmid and expressed as a fusion protein with the addition of 16 additional amino acids at the N terminus of E7. Expression and purification have been described previously (31).
Peptide Synthesis-Peptides were prepared by solid-phase synthesis using a double coupling protocol on the model 430A Applied Biosystems automated peptide synthesizer. Deprotection, purification and identity testing were as described previously (30). The E7(20-29) peptide sequence was TDLYCYEQLN-amide. The scrambled E7 peptide sequence was YNELCQYDL-amide. Lyophilized peptides were taken up in 0.1 M Tris-HCI, pH 7.5, and 1 mM dithiothreitol and the pH adjusted to neutrality.
Co-immune Precipitations-Radiolabeled pRB6O was co-immune precipitated with HPV-16 E7 protein using an antibody against E7. Reactions contained 25 p1 of reticulocyte-translated pRB6O and 0.15 pg of E7 diluted to 0.2 ml in LB buffer (25 mM HEPES, pH 7.2, 250 mM NaCl, 0.1% Triton X-100). Control reactions contained either no E7, 50 p~ E7(20-29) peptide, or 50 p~ scrambed E7 peptide. Reactions were incubated at 4 "C for 60 min after which time a polyclonal antibody raised to an E7 40-55 peptide was added and incubation continued 60 min. Antibody-bound E7-RB complexes were collected by adding 100 pI of a 1:10 slurry of protein A-Sepharose (Pharmacia) in LB buffer and incubating with gentle rocking for 40 min at 4 "C, followed by centrifugation to collect beads and washing 3 times in LB buffer. The washed Sepharose beads were resuspended in 60 pl of reducing SDS-polyacrylamide gel load buffer and heated to 90 "C for 10 min. Analysis of co-precipitated pRB6O was by SDS-polyacrylamide gel electrophoresis (SDS-PAGE) on 12% gels. The gels are fixed, treated with En3Hance (Du Pont-New England Nuclear), and visualized by autoradiography.
ELISA Assay-An ELISA assay to quantitate pRB6O binding to HPV-16 E7 protein was similar to that previously described (30). Briefly, reticulocyte lysate translations of 35S-labeled pRB6O and pRB6O mutants were quantitated for full-length protein by electrophoresis on SDS-PAGE followed by autoradiography and densitometer scanning of the pRB6O band. Equivalent volumes of reticulocyte translated pRB6O or mutants were sequentially diluted in LB buffer and added to a microtiter plate containing recombinant E7 protein.
After incubation and washing at 4 "C, two monoclonal RB antibodies, XZ104 and XZ107 (32), were added at a 1:lOO dilution in LB and incubated for 60 min at 4 "C. After washing, an alkaline phosphataseconjugated rabbit anti-mouse secondary antibody (Jackson Immunoresearch, West Grove, PA) was added at a 1:200 dilution and incubated at 4 "C for 60 min. After washing the wells, an alkaline phosphatase substrate (Kirkegaard & Perry Laboratories, Gaithersburg, MD) was added, and the reaction proceeded at 23 "C generating the colored product which was quantitated at 590 nM on a microtiter plate reader. Relative binding avidities as determined by the ELISA assay were normalized to the input full-length pRB6O as determined by densitometer scanning of an autoradiograph of the translation products electrophoresed on a reducing SDS-polyacrylamide gel.
DNA Binding Assay-The DNA binding assay was identical to that described previously (25). Briefly, reticulocyte lysate-translated pRB6O was diluted in HNMDT buffer (25 mM HEPES, 100 mM NaC1, 5 mM MgCl,, 1 mM dithiothreitol, 0.1% Triton X-100, pH 7.2) and passed over a 1.5-ml column of DNA-cellulose at 4 "C (Pharmacia). Flow-through was collected, and the column was subsequently washed with HNMDT buffer and eluted stepwise with HNMDT buffer containing 0.5 and 1.0 M NaCl. Fractions were analyzed by SDS-PAGE followed by fluorography. Gel loads of individual fractions (except wash) were normalized to collected volume so that the band intensities were representative of the relative percent of load which bound or flowed through.
pRB Antibody Immunoprecipitation-Equivalent quantities of 35Slabeled reticulocyte-translated pRB6O and pRB6O mutants were diluted 1:lO in LB buffer (final volume 250 pl) at 4, 23, or 30 "C, and anti-pRB monoclonal antibody XZ133 (32) was added at a final dilution of 1:100,000. The reactions were incubated at the desired temperatures for 40 min at which time 3 pl of undiluted rabbit antimouse antibody (Organon Teknika, West Chester, PA) was added and incubation continued for 40 min. Antibody-pRB complexes were precipitated by adding 100 pl of a 1:lO slurry of protein A-Sepharose to the reactions and incubating at the respective temperatures for an additional 40 min. Sepharose pellets were collected by centrifugation and washed with LB buffer at 4 "C. Analysis of immunoprecipitation pellets was as described above in the co-immune precipitation assay.

RESULTS
The RB protein binding pocket contains 8 cysteine residues (see Fig. 1). Four cysteines lie in domain A, 3 lie in domain B, and 1 cysteine is present in the linker region between domains A and B. Each of these cysteines was individually mutated to an alanine residue. The cysteine residue at position 853 in the C-terminal region of pRB6O lying outside the protein binding pocket was also changed to an alanine. Additionally, the cysteine residues at positions 666 and 706 in the binding pocket were changed to either serine or phenylalanine. To examine the effects of these substitutions on the binding properties of pRB60, each of the mutated RB genes was transcribed and translated in vitro to produce radiolabeled pRB6O proteins harboring specific cysteine substitutions (see "Materials and Methods").
The radiolabeled pRB6O wild type and mutant proteins were individually mixed with bacterially produced recombinant HPV-16 E7 protein (31) and immunoprecipitated using an anti-E7 antibody. As seen in Fig. 2, the translation products of each pRB6O species exhibited similar levels of radiolabel incorporation and product purity as evidenced by comparable SDS-PAGE gel analyses. In comparison to the wild type pRB6O immunoprecipitate, it is evident that alanine substitutions at cysteine residues 438, 489, 590, 712, and 853 had little or no effect on pRB binding to the E7 protein. By contrast, alanine substitutions at cysteine residues 407, 553, 666, and 706 produced pRB6O species with varying degrees of reduced binding to E7 protein. To better quantitate the E7 binding activity of these mutated pRB6O species, a separate series of experiments were performed using a pRB-HPV-16 E7 protein ELISA-style binding assay. In this assay, recom- The slowest migrating band represents pRB6O with other bands representing smaller translation products initiated at internal start sites. The wild type and mutated pRB6Os were co-immune-precipitated using an anti-E7 antibody in the presence of recombinant HPV-16 E7 (lanes A ) , co-precipitated in the absence of E7 (lanes B ) , coprecipitated with E7 in the presence of the E7(20-29) peptide (lanes C), or co-precipitated with E7 in the presence of the scrambled E7(20-29) peptide (lanes D). Only full-length pRB60 is co-precipitated by E7 and the E7 antibody. The two panels represent two separate experiments in which all reactions were performed in parallel, and, as such, individual mutants should only be compared with one another within experiment A or experiment B. binant E7 protein was bound to a solid matrix. Reticulocyte lysate-prepared pRB60 species were added to the matrix in solution. The unbound pRB6O was washed away, and the bound pRB60 was quantitated using an anti-pRB monoclonal antibody. The results of these assays are reported in Table I. In general, these studies are in agreement with the immunoprecipitation analyses. Alanine substitution at position 706 was most deleterious to E7 binding followed by similar substitutions at positions 666, 407, and 553. Substitutions of serines for cysteines at positions 666 and 706 and substitution of phenylalanine for cysteine at position 706 also impaired E7 binding. The phenylalanine substitution at position 706 had the most deleterious effect on E7 binding activity, while the serine substitution at this position was less deleterious. The alanine substitution had the least effect on E7 binding. In contrast to the relative effects of serine versus alanine substitutions at position 706, the alanine substitution at position 666 reduced E7 binding more than the serine substitution at the same position.
Substitution mutations for cysteine residues involved in hydrogen bonding or coordinate metal ion bonding might be expected to exhibit additive deleterious effects on E7 binding when incorporated into the same mutant protein. Alterna-

TABLE I Quuntitation of pRB6O-HPV-16 E7 interaction in ELISA-style
binding assays Reticulocyte translated radiolabeled pRB6O and pRB6O mutants were analyzed for their ability to bind recombinant HPV-16 E7 protein as described under "Materials and Methods." The data presented represent the average of at least two experiments. All assays were run in triplicate. The variation between triplicates was less than 10%. tively, substitution mutations for cysteine residues involved in disulfide bonds would not be expected to sum their effects on E7 binding once the possibility of forming a critical disulfide bond was eliminated by mutating either single residue from the cysteine pair comprising the disulfide bond. To determine whether multiple simultaneous cysteine substitutions would exhibit additive effects on pRB6O binding to E7 protein, mutant pRB6O species were created that combined two or more alanine substitutions. The 4 cysteine residues at positions 407, 553, 666, and 706 were initially chosen for analysis since single alanine substitutions at these positions showed the greatest effect on E7 binding activity. A pRB60 species with alanine substitutions at positions 407 and 553 was created, and a second pRB60 species with substitutions at positions 666 and 706 was also made since these pairs of cysteine residues exhibited similar levels of decreased E7 binding activity when examined as single substitution mutants. The pRB6O proteins corresponding to these substitution mutants were prepared and analyzed in the E7 ELISA binding assay. In both cases, the combination of two alanine substitutions in the same RB gene yielded pRB6O species with a greater loss of E7 binding activity than seen with either alanine substitution alone (see Table I). This result suggests that these residues do not participate in disulfide bonds in the pairwise combinations examined here. Lastly, a mutant pRB6O species was prepared combining all five cysteine to alanine substitutions within pRB6O that did not affect E7 binding activity. The cysteines at positions 438,489,590,712, and 853 were converted to alanines and the resulting pRB6O species was designated Cys-4 to indicate its retention of only 4 cysteine residues. The pRB60 protein generated from the Cys-4 mutant was analyzed in both immunoprecipitation and ELISA assays. As seen in Fig. 3 and Table I, combination of the five neutral alanine substitutions in the Cys-4 protein did not adversely affect pRB6O's E7 binding activity.

E7 protein species
The effect of each alanine substitution mutation on pRB6O's DNA binding activity was assessed using DNAcellulose columns. As seen in Fig. 4, wild type pRB6O bound to DNA-cellulose and was eluted with 0.5 M NaCl. By comparison to the wild type pRB60, alanine substitutions at positions 407, 553, 666, and 706 yielded pRB6O species with reduced DNA binding activity. Serine substitutions for cysteines at positions 666 and 706 and a phenylalanine substitution at position 706 also reduced pRB6O's DNA binding activity. Additional DNA binding studies were performed using pRB6O species harboring multiple alanine substitutions. The combination of alanine substitutions at positions 666 and 706 decreased DNA binding activity to a greater extent than occurred with either substitution alone. By contrast, the combination of alanine substitutions for cysteines at positions 438, 489, 590, 712, and 853 in the Cys-4 mutant had no apparent effect on pRB6O's ability to bind DNA. These results were similar to the effects of substitution mutations on the E7 binding activity of pRB6O described above. The same cysteine residues that were important for maintaining pRB6O's ability to bind E7 protein were also important for maintaining pRB6O's ability to bind DNA. Similarly, those substitution mutants which did not affect pRB6O binding to E7 protein also did not affect pRB60 binding to DNA.
The tight correlation between the E7 and DNA binding properties of the pRB6O substitution mutants suggests that the protein domains within pRB6O that are responsible for both of these biochemical properties are identical or exist in close proximity to one another. Alternatively, the correlation between the E7 protein and DNA binding properties of pRB6O may reflect a common response to conformational changes in pRB6O caused by these substitution mutations. To explore how these mutations might affect both the E7 protein and DNA binding properties of pRB60, a new series of immunoprecipitation studies was performed. These experiments were carried out using a conformation dependent anti-pRB monoclonal antibody XZ133 (32)  antibody. The experiment was conducted as described in Fig. 2. The reticulocyte lysate translation products of the WT pRB6O and Cys-4 pRB6O species are shown on the left. The relative levels of pRB6O coprecipitation in the presence or absence of E7 protein and E7 peptides are shown on the right. Lanes are labeled as follows: A, pRB6O plus E7; B, pRB6O minus E7; C, pRB6O plus E7 and E7(20-29) peptide; D, pRB6O plus E7 and scrambled E7(20-29) peptide.

FIG. 4.
Binding of pRB6O and pRB6O mutants to DNA. Reticulocyte translated 9-labeled pRB6O and substitution mutants were passed over DNAcellulose columns in HNMDT buffer at 4 "C and eluted as described under "Materials and Methods." The fractions collected were analyzed by SDS-PAGE and fluorography. The impaired binding ability of the mutants at this salt concentration is illustrated by increased amounts of full-length pRB6O in the flowthrough. Gel loads were normalized for collected total volume of the individual fractions to give a semi-quantitative comparison of pRB6O distribution between flow-through and DNA-bound fractions.
XZ133 antibody recognizes native RB proteins that are capable of binding viral transforming proteins, but does not bind to denatured RB proteins. As seen in Fig. 5, at 4 "C the XZ133 antibody effectively immunoprecipitated wild type pRB6O and the Ala-407, Ala-489, Ala-553, Ala-666, Ser-666, Ala-706, and the Cys-4 combination mutants. In contrast, the Ala-407 plus Ala-553, Ala-666 plus Ala-706, and the Phe-706 mutant proteins showed a reduced efficiency of precipitation relative to wild type pRB6O.
Since conformational flexibility can change with increasing temperature and altered flexibility might be expected to affect the conformation of pRB60, the immune precipitation reactions were also run at 23 and 30 "C. The lower panels in Fig.  5 show that several of the substitution mutants which behaved similarly to wild type pRB6O at 4 "C exhibited progressively poorer binding to the antibody with increasing temperature.
At 30 "C all of the mutants which showed altered E7 protein or DNA binding activity also bound less well to the conformation-dependent antibody. Again, the combinations of alanine substitutions at positions 407 plus 553 and 666 plus 706 yielded pRB6O species less capable of immunoprecipitation by XZ133 than found with the corresponding single-residue substitution mutants. As in the E7 protein and DNA binding assays the Phe-706 mutation was the most deleterious to antibody binding. The Ala-489 mutation and the Cys-4 combination mutations bound the antibody with the same efficiency as wild type pRB6O at all temperatures. A second conformation-sensitive monoclonal antibody, XZ104, was examined at 4 "C and yielded the same results as XZ133 (data not shown). In addition, a polyclonal rabbit antisera to pRB6O was examined. As expected, the polyclonal antisera, which presumably does not depend on pRB6O's conformation for its binding activity, precipitated all the mutant pRB60 species equally well. Taken together, these studies argue that substitution mutations at positions 407, 553, 666, and 706 in the pRB6O protein binding pocket alter the conformation of pRB6O. The progressive deleterious effect of increasing temperature on the binding of XZ133 to the four critical cysteine mutants suggests deterioration of pRB6O's conformational stability at higher temperatures. This result supports the hypothesis that the 4 critical cysteines at positions 407, 553, 666, and 706 participate in maintaining the conformational integrity of the pRB6O protein binding pocket. Immune precipitation of pRB6O and pRB6O mutants with a conformationally sensitive antibody. Wild type pRB6O and the pRB6O substitution mutants were immune-precipitated using the conformationally sensitive antibody XZ133 at three different temperatures as described under "Materials and Methods." Equivalent amounts of radiolabeled full-length pRB6O were used in each reaction. The collected protein A-Sepharose pellets were analyzed for the level of pRB6O bound to the antibody using SDS-PAGE analysis followed by fluorography. position 666, and alanine, serine, or phenylalanine substitutions at position 706 all adversely affected pRB6O's protein and DNA binding activities. Amino acid changes at position 706 exhibited the strongest deleterious effects on RB6O's ability to bind E7 protein or DNA. Interestingly, position 706 is also the site of a naturally occurring phenylalanine substitution mutation associated with a small cell lung cancer (26). Single-residue mutations at position 666 exhibited the next most deleterious effects on binding activities, while substitutions at positions 407 and 553 showed lesser degrees of impairment. By contrast, alanine substitutions at positions 438, 489,590,712, and 853 caused no detectable change in pRB6O's binding properties. It should be noted that these neutral mutations were scattered throughout pRB60 while the four delterious mutations were all found in the A and B domains (see Fig. 1). These observations are consistent with the hypothesis that the pRB binding pocket is a critical region that contributes to both the protein and DNA binding properties of RB proteins. Moreover, these studies suggest that the cysteine residues at positions 407, 553, 666, and particularly 706 play an essential role in maintaining these biochemical properties.

DISCUSSION
It is important to note that the E7 protein and DNA binding properties of pRB6O were affected similarly by the substitution mutants examined in this study. The five neutral mutations failed to alter either E7 protein binding or DNA-cellulose binding while the four deleterious mutations impaired both of these activities. Moreover, those mutations which most severely affected pRB6O's E7 protein binding activity, substitutions at positions 666 and 706, also exhibited maximal impairment of DNA binding. This apparent linkage of the E7 protein and DNA binding properties of pRB6O was not entirely unexpected. We previously showed that E7 protein binding to pRB60 impaired the ability of pRB6O to bind DNA (25). Those experiments suggested that pRB6O's E7 protein and DNA binding properties were influenced by the same or adjacent regions of the RB protein. The current study indicates that changes in specific cysteine residues in the binding pocket of pRB6O affect both of these biochemical properties in parallel. Two mechanisms can be hypothesized to explain the parallel fall in pRB6O's binding activities associated with the substitution mutations at positions 407,553,666, and 706. First, the cysteine residues normally present at these positions may serve as direct contact points between pRB6O and both E7 protein and DNA molecules. Second, these cysteine residues may help to form or maintain the proper conformation of pRB6O needed to bind E7 protein and DNA. Our data supports the latter hypothesis since it appears unlikely that the pRB6O cysteines at positions 407,553,666, and 706 would form contacts with both E7 protein and DNA while the nearby cysteines at positions 438, 489, 590, and 712 failed to make contact with either E7 protein or DNA. Moreover, the decrease in conformation-specific antibody recognition of pRB60 associated with the substitution mutations at positions 407,553,666, and 706 and the temperature sensitivity of this recognition suggests that these mutations led to a change in pRB6O's conformational structure. This interpretation of our immunoprecipitation analyses is supported by the fact that the antibody used in these studies, XZ133, is dependent upon the conformation of pRB6O in order to recognize and immunoprecipitate RB proteins (32). Moreover, the differential effect of temperature on antibody binding to the critical substitution mutants suggests that the native conformation of pRB6O becomes labile at higher temperatures. The epitope recognition site for this antibody has been mapped to RB protein residues 444-535 or 620-665 (32). These segments of the RB protein do not contain any of the critical cysteine residues identified in this study. Therefore, the antibody's impaired ability to bind the critical cysteine mutants cannot be attributed to the loss of contact site residues within the A and B domains of the binding pocket.
The current study clearly demonstrates that the cysteine residues at positions 407, 553, 666, and 706 are important amino acids in maintaining pRB6O's biochemical functions. Examination of the amino acid sequence of the related adenovirus E1A and HPV-16 E7 binding protein, p107 (28), also suggests that these residues are important. While the overall amino acid conservation between pRB105 and p107 is 34%,4 out of 8 cysteine residues-407,489,666, and 706-within the pRB binding pocket are conserved in p107. Three of these 4 residues (407, 666, and 706) are important in maintaining pRB6O's biochemical binding properties. By contrast, only 1 of the 4 cysteine residues that are not conserved between the pRB binding pocket and p107 (553,438,590, and 712) proved to be important for maintaining binding activity. Substitution of alanine for cysteine at position 553 reduced pRB6O's binding properties. This correlation between the conservation of cysteine residues in pRB and p107 and the requirement of these residues for maintaining pRB6O's binding properties reinforces the hypothesis that cysteine residues play a critical role in cellular proteins that bind the E1A and E7 viral oncoproteins.
Cysteine residues are known to serve at least three structural functions in proteins: formation of disulfide bonds, hydrogen bonding, and coordinate bond formation with metal ions. The results presented here do not conclusively implicate or eliminate any of these functions as possible mechanisms underlying the function of the cysteines at positions 407,553, 666, and 706. However, disulfide formation appears to be an unlikely mechanism involving these residues both because the internal milieu surrounding intracellular proteins does not favor disulfide bond formation, and because the pairwise substitution mutations examined in this study did not support this hypothesis. Additional biophysical studies will be needed to determine whether the critical cysteine residues in pRB6O's protein binding pocket are simply involved in maintaining the proper conformation of RB proteins or participate in other biochemical functions.