The anti-apoptotic activity of XIAP is retained upon mutation of both the caspase 3- and caspase 9-interacting sites

The X-linked mammalian inhibitor of apoptosis protein (XIAP) has been shown to bind several partners. These partners include caspase 3, caspase 9, DIABLO/Smac, HtrA2/Omi, TAB1, the bone morphogenetic protein receptor, and a presumptive E2 ubiquitin-conjugating enzyme. In addition, we show here that XIAP can bind to itself. To determine which of these interactions are required for it to inhibit apoptosis, we generated point mutant XIAP proteins and correlated their ability to bind other proteins with their ability to inhibit apoptosis. partial differential RING point mutants of XIAP were as competent as their full-length counterparts in inhibiting apoptosis, although impaired in their ability to oligomerize with full-length XIAP. Triple point mutants, unable to bind caspase 9, caspase 3, and DIABLO/HtrA2/Omi, were completely ineffectual in inhibiting apoptosis. However, point mutants that had lost the ability to inhibit caspase 9 and caspase 3 but retained the ability to inhibit DIABLO were still able to inhibit apoptosis, demonstrating that IAP antagonism is required for apoptosis to proceed following UV irradiation.

HtrA2 promotes cell death through its serine protease activity and its ability to antagonize inhibitor of apoptosis proteins

Inhibitor of apoptosis (IAP) proteins inhibit caspases, a function counteracted by IAP antagonists, insect Grim, HID, and Reaper and mammalian DIABLO/Smac. We now demonstrate that HtrA2, a mammalian homologue of the Escherichia coli heat shock-inducible protein HtrA, can bind to MIHA/XIAP, MIHB, and baculoviral OpIAP but not survivin. Although produced as a 50-kDa protein, HtrA2 is processed to yield an active serine protease with an N terminus similar to that of Grim, Reaper, HID, and DIABLO/Smac that mediates its interaction with XIAP. HtrA2 is largely membrane-associated in healthy cells, with a significant proportion observed within the mitochondria, but in response to UV irradiation, HtrA2 shifts into the cytosol, where it can interact with IAPs. HtrA2 can, like DIABLO/Smac, prevent XIAP inhibition of active caspase 3 in vitro and is able to counteract XIAP protection of mammalian NT2 cells against UV-induced cell death. The proapoptotic activity of HtrA2 in vivo involves both IAP binding and serine protease activity. Mutations of either the N-terminal alanine of mature HtrA2 essential for IAP interaction or the catalytic serine residue reduces the ability of HtrA2 to promote cell death, whereas a complete loss in proapoptotic activity is observed when both sites are mutated.

Relative dominance of epitope-specific cytotoxic T-lymphocyte responses in human immunodeficiency virus type 1-infected persons with shared HLA alleles

Cytotoxic T lymphocytes (CTL) target multiple epitopes in human immunodeficiency virus (HIV)-infected persons, and are thought to influence the viral set point. The extent to which HLA class I allele expression predicts the epitopes targeted has not been determined, nor have the relative contributions of responses restricted by different class I alleles within a given individual. In this study, we performed a detailed analysis of the CTL response to optimally defined CTL epitopes restricted by HLA class I A and B alleles in individuals who coexpressed HLA A2, A3, and B7. The eight HIV-1-infected subjects studied included two subjects with acute HIV infection, five subjects with chronic HIV infection, and one long-term nonprogressor. Responses were heterogeneous with respect to breadth and magnitude of CTL responses in individuals of the same HLA type. Of the 27 tested epitopes that are presented by A2, A3, and B7, 25 were targeted by at least one person. However, there was wide variation in the number of epitopes targeted, ranging from 2 to 17. The A2-restricted CTL response, which has been most extensively studied in infected persons, was found to be narrowly directed in most individuals, and in no cases was it the dominant contributor to the total HIV-1-specific CTL response. These results indicate that HLA type alone does not predict CTL responses and that numerous potential epitopes may not be targeted by CTL in a given individual. These data also provide a rationale for boosting both the breadth and the magnitude of HIV-1-specific CTL responses by immunotherapy in persons with chronic HIV-1 infection.

Direct inhibition of caspase 3 is dispensable for the anti-apoptotic activity of XIAP

XIAP is a mammalian inhibitor of apoptosis protein (IAP). To determine residues within the second baculoviral IAP repeat (BIR2) required for inhibition of caspase 3, we screened a library of BIR2 mutants for loss of the ability to inhibit caspase 3 toxicity in the yeast Schizosaccharomyces pombe. Four of the mutations, not predicted to affect the structure of the BIR fold, clustered together on the N-terminal region that flanks BIR2, suggesting that this is a site of interaction with caspase 3. Introduction of these mutations into full-length XIAP reduced caspase 3 inhibitory activity up to 500-fold, but did not affect its ability to inhibit caspase 9 or interact with the IAP antagonist DIABLO. Furthermore, these mutants retained full ability to inhibit apoptosis in transfected cells, demonstrating that although XIAP is able to inhibit caspase 3, this activity is dispensable for inhibition of apoptosis by XIAP in vivo.

Crystal structure of the amino-terminal coiled-coil domain of the APC tumor suppressor

Coiled coils serve as dimerization domains for a wide variety of proteins, including the medically important oligomeric tumor suppressor protein, APC. Mutations in the APC gene are associated with an inherited susceptibility to colon cancer and with approximately 75 % of sporadic colorectal tumors. To define the basis for APC pairing and to explore the anatomy of dimeric coiled coils, we determined the 2.4 A resolution X-ray crystal structure of the N-terminal dimerization domain of APC. The peptide APC-55, encompassing the heptad repeats in APC residues 2-55, primarily forms an alpha-helical, coiled-coil dimer with newly observed core packing features. Correlated asymmetric packing of four core residues in distinct, standard rotamers is associated with a small shift in the helix register. At the C terminus, the helices splay apart and interact with a symmetry-related dimer in the crystal to form a short, anti-parallel, four-helix bundle. N-terminal fraying and C-terminal splaying of the helices, as well as the asymmetry and helix register shift describe unprecedented dynamic excursions of coiled coils. The low stability of APC-55 and divergence from the expected coiled-coil fold support the suggestion that the APC dimerization domain may extend beyond the first 55 residues.

Solution structure and mutagenesis of the caspase recruitment domain (CARD) from Apaf-1

Activation of procaspase-9, a key component of the apoptosis mechanism, requires the interaction of its caspase recruitment domain (CARD) with the CARD in the adaptor protein Apaf-1. Using nuclear magnetic resonance spectroscopy and mutagenesis we have determined the structure of the CARD from Apaf-1 and the residues important for binding the CARD in procaspase-9. Apaf-1's CARD contains seven short alpha-helices with the core six helices arranged in an antiparallel manner. Residues in helix 2 have a central role in mediating interaction with procaspase-9 CARD. This interaction surface is distinct from that proposed based on the structure of the CARD from RAIDD, but is coincident with that of the structurally similar FADD death effector domain and the Apaf-1 CARD interface identified by crystallographic studies.

Solution structure of a baculoviral inhibitor of apoptosis (IAP) repeat

Members of the inhibitor of apoptosis (IAP) family of proteins are able to inhibit cell death following viral infection, during development or in cell lines in vitro. All IAP proteins bear one or more baculoviral IAP repeats (BIRs). Here we describe the solution structure of the third BIR domain from the mammalian IAP homolog B (MIHB/c-IAP-1). The BIR domain has a novel fold that is stabilized by zinc tetrahedrally coordinated by one histidine and three cysteine residues. The structure consists of a series of short alpha-helices and turns with the zinc packed in an unusually hydrophobic environment created by residues that are highly conserved among all BIRs.

Mutation of arginine 121 in lactoferrin destabilizes iron binding by disruption of anion binding: crystal structures of R121S and R121E mutants

A conserved arginine residue helps to form the synergistic anion binding site in transferrins. To probe the importance of this residue for anion binding and iron binding, Arg 121 has been mutated to Ser and Glu in N-terminal half-molecule of human lactoferrin. The two mutants, R121S and R121E, have been expressed, purified, and crystallized. Their three-dimensional structures have been determined by X-ray diffraction at 2.3 and 2.5 A resolution, respectively. The structures were determined by molecular replacement and were refined by restrained least squares methods to final R values of 0.185 and 0.204. Both mutants still bind iron but with decreased stability. The crystal structures show that destabilization of iron binding probably results from disruption of the anion binding site; mutation of Arg 121 removes one wall of the anion binding pocket and causes the synergistic carbonate ion to be displaced 0.5 A from its position in the wild-type protein. In the process it becomes partially detached from the helix N-terminus that forms the rest of the anion binding site.

Altered domain closure and iron binding in transferrins: the crystal structure of the Asp60Ser mutant of the amino-terminal half-molecule of human lactoferrin

The crystal structure of a site-specific mutant of the N-terminal half-molecule of human lactoferrin, Lf(N), in which the iron ligand Asp60 has been mutated to Ser, has been determined at 2.05 A resolution in order to determine the effects of the mutation on iron binding and domain closure. Yellow monoclinic crystals of the D60S mutant, in its iron-bound form, were prepared, and have unit cell dimensions a = 110.2 A, b = 57.0 A, c = 55.2 A, beta = 97.6 degrees, space group C2, with one molecule of 333 residues in the asymmetric unit. The structure was determined by molecular replacement, using the wild-type Lf(N) as search model, and was refined by restrained least-squares methods. The final model, comprising 2451 protein atoms (from residues 2 to 315) one Fe3+ and one CO2-(3), and 107 water molecules, gives an R-factor of 0.175 for all data in the resolution range 20.0 to 2.05 A. The model conforms well with standard geometry, having root-mean-square deviations of 0.014 A and 1.2 degrees from standard bond lengths and angles. The structure of the D60S mutant deviates in two important respects from the parent Lf(N) molecule. At the mutation site the Ser side-chain neither binds to the iron atom nor makes any interdomain contact as the substituted Asp does; instead a water molecule fills the iron coordination site and participates in interdomain hydrogen bonding. The domain closure is also changed, with the D60S mutant having a more closed conformation. Consideration of crystal packing suggests that the altered domain closure is a genuine molecular property but both the iron coordination and interdomain contacts are consistent with weakened iron binding in the mutant. The implications for iron binding in transferrins generally are discussed.

Enzymatic deglycosylation as a tool for crystallization of mamalian binding proteins

Enzymatic deglycosylation has been used in attempts to crystallize several glycoproteins with the aim of overcoming the problems resulting from heterogeneity and flexibility of the attached glycan chains. An endoglycosidase preparation from Flavobacterium meningosepticum, comprising the enzymes endo F and PNGase-F, was used in experiments on the mammalian binding proteins lactoferrin and haemopexin. Significant differences were found in the susceptibility of different proteins to deglycosylation. For human lactoferrin (Lf) and its recombinant N-terminal half-molecule (Lf(N)), deglycosylation was rapid and complete, and was essential for obtaining high-quality crystals of both apo-Lf and Lf(N); for bovine Lf, however, complete deglycosylation did not occur. Similarly, for rabbit haemopexin the carbohydrate chain on the C-terminal domain was easily removed, but the three chains on the N-terminal domain proved more resistant and their removal led to some fragmentation of the protein. Nevertheless, this approach provided the only means of crystallizing the C-terminal domain and is likely to be useful for other glycoproteins.

Three-dimensional structure of lactoferrin in various functional states

The three-dimensional structures of various forms of lactoferrin, determined by high resolution crystallographic studies, have been compared in order to determine the relationship between structure and biological function. These comparisons include human apo and diferric lactoferrins, metal and anion substituted lactoferrins, the N-terminal half molecule of human lactoferrin, and bovine diferric lactoferrin. The structures themselves define the nature and location of the iron binding sites and allow anti-bacterial and putative receptor-binding regions to be mapped on to the molecular surface. The structural comparisons show that small internal adjustments can allow the accommodation of different metals and anions without altering the overall molecular structure, whereas large-scale conformational changes are associated with metal binding and release, and smaller, but significant, movements accompany species variations. The results also focus on differences in flexibility between the two lobes, and on the importance of interactions in the inter-lobe region in modulating iron release from the N-lobe and in possibly enabling binding at one site to be signalled to the other.

Structure of the recombinant N-terminal lobe of human lactoferrin at 2.0 A resolution

The three-dimensional structure of the N-terminal half-molecule of human lactoferrin, LfN, prepared by recombinant DNA methods, has been determined by X-ray crystallography at 2.0 A resolution. The protein is in its iron-bound form and is deglycosylated. X-ray diffraction data were obtained by diffractometry to 3.2 A resolution and synchrotron data collection, using Weissenberg photography with imaging plates, to 1.8 A resolution. The structure was solved by molecular replacement, using the N-lobe of native diferric human lactoferrin (Lf) as search model. Restrained least squares refinement (program TNT) has resulted in a model structure with an R-factor of 0.184 for all data 34,180 (reflections) in the resolution range 8.0 to 2.0 A. The model comprises 2490 protein atoms (residues 4 to 327), 1 Fe3+, 1 CO3(2-) and 180 solvent molecules, all regarded as water. The structure of LfN is essentially the same as that of the N-lobe of intact Lf, being folded into two similar alpha/beta domains, with the Fe3+ and CO3(2-) bound in a specific site in the interdomain cleft. These details are not affected by either deglycosylation or expression in a non-native system. At the C terminus, however, the conformation of residues 321 to 333 is changed. Whereas in Lf residues 321 to 332 form a helix crossing between the domains at the back of the iron site, in LfN residues 321 to 326 have an extended conformation, forming a third interdomain beta-strand, and residues 328 to 333 appear disordered. The conformational change is attributed to the loss of stabilizing interactions from the C-lobe and is mediated by two Gly residues, at positions 321 and 323. It is further proposed that the conformational change is responsible for the more facile iron release properties of LfN, by its effect on the hinge mechanism and increased solvent exposure of residues near the back of the iron site. Other details of the polypeptide chain conformation and the binding site have also been analysed. Two cis-proline residues are found at positions 71 and 142. The bidentate binding of the CO3(2-) to the metal ion is unambiguous, and a network of hydrogen bonds in and around the binding site links the two domains. Clearly-defined amino-aromatic hydrogen bonds are found for Arg210, near the metal site, and some 31 internal water molecules have been identified, 15 of them in essentially discrete sites, and 16 in a cluster filling a cavity in the interdomain cleft.

Preliminary crystallographic studies of the amino terminal half of human lactoferrin in its iron-saturated and iron-free forms

The amino terminal half of human lactoferrin (LfN) produced from transfected baby hamster kidney cells has been crystallized in its iron-saturated and iron-free forms. The crystals of glycosylated LfN and deglycosylated LfN are monoclinic, space group C2, with cell dimensions a = 133.0 A, b = 58.3 A, c = 58.3 A, alpha = 90.0 degrees, beta = 114.7 degrees, gamma = 90.0 degrees, and one molecule per asymmetric unit. Crystals of apo LfN have also been prepared using deglycosylated protein. These crystals are tetragonal, space group P4(1)2(1)2 (or P4(3)2(1)2), with cell dimensions of a = b = 58.4 A and c = 217.2 A and one molecule per asymmetric unit. Both the iron-saturated and the iron-free crystals are suitable for high resolution X-ray analysis.

Studies of the N-terminal half of human lactoferrin produced from the cloned cDNA demonstrate that interlobe interactions modulate iron release

The factors influencing iron binding and release by lactoferrin have been addressed by comparison of the native full length molecule (Lf) with the N-terminal half of human lactoferrin (LfN) produced from the cloned cDNA expressed in baby hamster kidney (BHK) cells. The coding sequences for LfN were inserted into the expression vector pNUT between the metallothionein promoter and the human growth hormone transcription termination sequences. Transformed BHK cells were grown in roller bottles where concentrations of LfN as high as 35 mg/liter were obtained. The pure protein, produced by the transformed BHK cells, was characterized by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, protein blotting and immunodetection, N-terminal sequence analysis, UV-visible spectroscopy, electron spin resonance spectroscopy, and measurements of metal binding and release. By these criteria LfN was found to be correctly processed, glycosylated, and able to bind iron reversibly. Both UV-visible and electron spin resonance spectra of the half molecule were very similar to those of native lactoferrin and the full length lactoferrin produced in BHK cells, but there were marked differences in the pH at which iron release occurred. Iron release from LfN occurs in the pH range 6.0-4.0, compared with 4.0-2.5 for native lactoferrin and 6.2-4.0 for transferrin. These results suggest that the more facile release of iron from LfN compared with native lactoferrin results from the absence of stabilizing contacts between the N- and C-terminal halves and that the characteristic difference in pH stability between lactoferrins and transferrins is due primarily to differences in these interactions.

Prognostic factors for local recurrence, metastasis, and survival rates in squamous cell carcinoma of the skin, ear, and lip. Implications for treatment modality selection

We reviewed all studies since 1940 on the prognosis of squamous cell carcinoma (SCC) of the skin and lip. The following variables are correlated with local recurrence and metastatic rates: (1) treatment modality, (2) prior treatment, (3) location, (4) size, (5) depth, (6) histologic differentiation, (7) histologic evidence of perineural involvement, (8) precipitating factors other than ultraviolet light, and (9) host immunosuppression. Local recurrences occur less frequently when SCC is treated by Mohs micrographic surgery. This local recurrence rate differential in favor of Mohs micrographic surgery holds true for primary SCC of the skin and lip (3.1% vs 10.9%), for ear SCC (5.3% vs 18.7%), for locally recurrent (previously treated) SCC (10% vs 23.3%), for SCC with perineural involvement (0% vs 47%), for SCC of size greater than 2 cm (25.2% vs 41.7%), and for SCC that is poorly differentiated (32.6% vs 53.6%).