Cytotoxic T lymphocyte response to minor H-43a alloantigen in H-43b mice. Privileged H-2Kb restriction to the response is not due to immunodominance or epistatic effect but due to Ir gene function of H-2Kb itself

Previous study demonstrated that anti-H-43a cytotoxic T lymphocyte (CTL) response of H-43b CWB (H-2b) stain carrying non-major histocompatability complex (MHC) genes of C3H and F1 strains raised by crossing CWB with various H-43b strains was restricted exclusively by self H-2Kb (Kb). In the present study, newly produced C3W strain (H-2k, H-43b), which is H-43-congenic to C3H/HeN (H-2k, H-43a), was used as H-43b mice, and possibility of immunodominance of Kb was examined. No anti-H-43a CTL response could be induced in C3W strain and F1 strains raised by crossing C3W with other H-43b strains not carrying Kb. Thus, the possibility of immunodominance of Kb over the other MHC class I alleles could not be supported. We also examined possibility of epistatic effect of I region genes and non-MHC genes on the Kb restriction. (C3W x C57BL/6)F1(I-Ak/b) and (C3W x B6.CH-2bm12)F1(I-Ak/bm12)mice showed equally anti-H-43a CTL response restricted exclusively by self Kb, and (C3W x B10.MBR)F1(Ik/k) mice also showed anti-H-43a CTL response restricted solely by self Kb. Cold target competition experiments demonstrated that H-43b C57BL/10 or A.BY mice, which do not have non-MHC genes of C3H mounted anti-H-43a CTL response restricted solely by self Kb. Thus, no relation of I region genes or non-MHC genes to the Kb restriction was shown. All the results indicate that H-43b mouse strains, including F1, can not achieve anti-H-43a CTL response unless they carry Kb allele. Notably, (C3W x C57BL/6)F1 mice mounted self Kb-restricted anti-H-43a CTL response, whereas (C3W x B6.CH-2bm1)F1 mice carrying mutated Kb could not mount anti-H-43a CTL response at all. These findings indicate strongly that Kb itself is classical Ir gene of anti-H-43a CTL response and directs self Kb restriction of the response.     

Ir gene control of the cytotoxic T lymphocyte response to Sendai virus: H-2k mice are low responders to Sendai

H-2k mice generate a secondary in vitro cytotoxic T lymphocyte response to Sendai virus 20- to 100-fold weaker than those of other haplotypes tested (H-2b,d,q,s). This immune response defect maps to both H-2K and H-2D. H-2k x H-2d F1 mice (responder x nonresponder) only lyse targets that have the d allele at H-2K and/or H-2D. H-2k targets are equally lysable with anti-Sendai antibody. Furthermore, H-2k mice demonstrate normal antibody and T cell proliferation responses to Sendai virus. The Ir gene defect therefore appears to be limited to the generation of the cytotoxic T lymphocytes.     

Resistance to H-2-restricted but not to allo-H2-specific graft and cytotoxic T lymphocyte responses in lymphoma mutant

The lymphoma mutant RMA-S escaped graft rejection after transplantation over a minor histocompatibility barrier, whereas it was rejected in H-2 allogeneic mice. The parental control line was rejected in both situations. The mutant, which had been selected against MHC class I molecules retained 5 to 10% of the wild-type H-2Db, Kb, and beta 2-microglobulin expression on the cell surface. It remained sensitive to allo-H-2b CTL in vitro, but was completely resistant to minor histocompatibility antigen-specific, H-2b-restricted CTL. It was equally resistant to other H-2b-restricted responses against internally derived Ag, such as tumor-specific CTL or a CTL clone specific for the influenza virus nucleoprotein. The results indicate a target cell defect that selectively abolishes the sensitivity to H-2-restricted CTL directed against internally processed Ag. This appears sufficient to shift the transplantation response over a minor histocompatibility Ag barrier from rejection to acceptance. There are two possible explanations for the results: 1) a block in the MHC class I-directed pathway for internal Ag processing, and 2) subthreshold H-2/Ag ligand density in relation to triggering requirements of restricted CTL. Regardless of the type of defect, the results demonstrate a difference between allo-H-2-specific and H-2-restricted CTL recognition at the level of the target cell.     

Anti-adenovirus type 5 cytotoxic T lymphocytes: immunodominant epitopes are encoded by the E1A gene.

Virus specific, major histocompatibility complex-restricted, cytotoxic T lymphocytes (CTL) generated in Fischer strain rats infected with human adenovirus type 5 (Ad5) were found to recognize antigenic determinants encoded within the Ad5 early region 1A (E1A) gene. Preliminary mapping studies suggest that the E1A CTL epitopes are encoded within the regions between bp 625 to 810 and 916 to 974 in the first exon of this gene. These epitope-coding regions occur within subregions of E1A that are conserved functionally, and to some extent structurally (approximately 50% sequence homology), among adenoviruses of different groups. Nevertheless, Ad5-specific CTL lysed only targets infected with adenoviruses of the same group (group C; e.g., Ad2) and not targets infected with adenoviruses of different groups (groups A, B, and E). These results suggest that virus-specific CTL may limit adenoviral dissemination by destroying virus-infected cells at an early stage in the viral replicative cycle, during E1A gene expression. Expression of other adenovirus genes does not appear to be required to target infected cells for elimination by CTL.     

A karyophilic signal sequence in adenovirus type 5 E1A is functional in Xenopus oocytes but not in somatic cells.

The carboxy-terminal five amino acids of the adenovirus type 5 E1A gene product are necessary and sufficient for this protein to become localized in the nuclei of somatic cells. In this report, we demonstrate that E1A contains a second nuclear localization signal, which resides within residues 140 to 185 and which functions in Xenopus oocytes but not Xenopus or mammalian somatic cells. These data demonstrate the cell-type-specific utilization of a nuclear localization signal.     

Mechanism of nonresponsiveness to AKR/Gross leukemia virus in AKR.H-2b:Fv-1b mice. An analysis of precursor cytotoxic T lymphocyte frequencies in young versus moderately aged mice

As young adult AKR.H-2b:Fv-1b mice reach about 9 wk of age, they begin to develop a nonresponsiveness to AKR/Gross leukemia virus. Unlike young mice that are responders, moderately aged AKR.H-2b:Fv-1b mice, after immunization and secondary in vitro restimulation in bulk culture with AKR/Gross virus induced tumors, can not generate anti-AKR/Gross virus-specific CTL. The mechanism of conversion to nonresponsiveness in moderately aged AKR.H-2b:Fv-1b mice is not understood, but it is correlated with increased expression of endogenous ecotropic viral antigens. Our present investigation focuses on determining the frequency of anti-AKR/Gross virus precursor CTL in AKR.H-2b:Fv-1b mice as a function of age. This was achieved by performing limiting dilution cultures of immune spleen cells obtained from young and moderately aged AKR.H-2b:Fv-1b mice. Although spleen cells obtained from immune moderately aged mice can not differentiate in bulk cultures into anti-AKR/Gross virus-specific CTL, there was no evidence of substantially decreased frequencies of virus-specific precursor CTL, relative to precursor CTL frequencies observed in young responder AKR.H-2b:Fv-1b mice.     

Inability of mice to generate cytotoxic T lymphocytes to vesicular stomatitis virus restricted to H-2Kk or H-2Dk

H-2k mice are unable to generate cytotoxic T lymphocytes (CTL) to vesicular stomatitis virus (VSV). This apparent unresponsiveness is found for both major serotypes of VSV, VSV-Indiana and VSV-New Jersey. CTL unresponsiveness occurs despite the ability of H-2k mice to generate a humoral immune response against VSV that is comparable to that found in responder (H-2b and H-2a) strains. All H-2k mice regardless of background genes, including various Ig allotypes, were found to be nonresponders. H-2k-linked unresponsiveness mapped to both H-2Kk and H-2Dk and occurred despite the presence of responder alleles in (responder x nonresponder)F1 mice. The unresponsiveness cannot be attributed to an inability of VSV-infected H-2k target cells to express viral surface antigens of H-2 molecules. Further, unresponsiveness cannot be overcome by using secondary stimulation in vivo or in vitro. H-2k-linked unresponsiveness does not appear to be due to suppression, and no complementation has been found in various (nonresponder x nonresponder)F1 mice. Thus unresponsiveness to VSV in association with H-2Kk or H-2Dk appears to represent an extensive defect of immune responsiveness that probably occurs because VSV is not a natural mouse pathogen.     

Cytotoxic T lymphocyte response to minor H-42a alloantigen in H-42b mice: clonal inactivation of the precursor cytotoxic T lymphocytes by veto-like spleen cells that express the H-42a antigen

When (B10.BR X CWB)F1 (BWF1; H-2k/b) mice carrying the H-42b allele at the minor H-42 locus were injected with H-42a C3H.SW (CSW; H-2b) or C3H (H-2k) spleen cells (SC), self-H-2Kb restricted anti-H-42a pCTL in the BWF1 recipients were primed and differentiated to anti-H-42a CTL after in vitro stimulation with (B10.BR X CSW)F1 (BSF1; H-2k/b, H-42b/a) SC. In contrast, anti-H-42a pCTL in H-42b mice were inactivated by injection with H-42-congenic H-42a SC, and stable anti-H-42a CTL tolerance was induced. Preference of H-2Kb restriction of anti-H-42a CTL was strict, and self-H-2Kb-restricted anti-H-42a CTL did not lyse target cells carrying H-42a antigen in the context of H-2Kbm1. Involvement of suppressor cells in the anti-H-42a CTL tolerance was ruled out by the present cell transfer study and the previous cell-mixing in vitro study. Notably, treatment with anti-Thy-1.2 antibody (Ab) plus complement (C) wiped out the ability of CSW SC in the priming of anti-H-42a pCTL of BWF1 mice but left that of C3H SC unaffected, and injection of the anti-Thy-1.2 Ab plus C-treated CSW SC induced anti-H-42a CTL tolerance in the BWF1 recipients. Furthermore, H-42a/b, I-Ab/bm12 [CSW X B6.CH-2bm12 (bm12)]F1 SC could not prime anti-H-42a pCTL in H-42b, I-Ab (CWB X B6)F1 recipients, whereas H-42a/b, I-Ab (CSW X B6)F1 SC primed anti-H-42a pCTL in H-42b, I-Ab/bm12 (CWB X bm12)F1 recipients. The unresponsiveness of anti-H-42a pCTL in H-42b mice to H-42-congenic H-42a SC was sometimes corrected by immunization of H-42b female mice with H-42-congenic H-42a male SC. Taking all of the results together, we propose the following. Unresponsiveness of anti-H-42a pCTL in H-42b mice to H-42-congenic H-42a SC is caused by "veto cells" contained in the antigenic H-42a SC. Anti-H-42a pCTL in the H-42b recipients directly interacting with H-42-congenic H-42a SC, which bear H-42a antigen and H-2Kb restriction element, are inactivated or vetoed.(ABSTRACT TRUNCATED AT 400 WORDS)     

Msx1 is close but not allelic to either Hm or Hx on mouse Chromosome 5

The Msx1 homeobox locus has been mapped in relation to the mutations hammer-toe (Hm) and hemimelic extra toes (Hx). Msx1 is expressed in the developing limb, while limb development is affected by the Hm and Hx mutations. Hm and Hx are very tightly linked loci. In interspecific crosses, the segregation of either mutation was followed in relation to polymorphic alleles of Msx1, Il6, and En2, to give a fine map around the mutant loci. Our results show that Msx1 is not allelic to either of the mutations, but is located about 3 cM from them. Il6 did not recombine with either Hm or Hx and, therefore, provides a point of access for the analysis of these mutations at the molecular level.

---

     

Cutting edge: myelin basic protein-specific cytotoxic T cell tolerance is maintained in vivo by a single dominant epitope in H-2k mice.

Multiple sclerosis (MS) is believed to be an autoimmune disease mediated by T cells specific for CNS Ags. MS lesions contain both CD4+ and CD8+ T lymphocytes. The contribution of CD4+ T cells to CNS autoimmune disease has been extensively studied in an animal model of MS, experimental autoimmune encephalomyelitis. However, little is known about the role of autoreactive CD8+ cytotoxic T cells in MS or experimental autoimmune encephalomyelitis. We demonstrate here that myelin basic protein (MBP) is processed in vivo by the MHC class I pathway leading to a MBP79-87/Kk complex. The recognition of this complex by MBP-specific cytotoxic T cells leads to a high degree of tolerance in vivo. This study is the first to show that the pool of self-reactive lymphocytes specific for MBP contain MHC class I-restricted T cells whose response is regulated in vivo by the induction of tolerance.     

A monoclonal antibody to viral glycoprotein blocks virus-immune effector T cells operating at H-2Dd but not at H-2Kd1.

A particular monoclonal antibody that binds to the influenza virus HA molecule inhibits HA-specific thymus-derived lymphocytes mediating cytotoxicity in the context of H-2Dd but not of H-2Kd. Another monoclonal antibody blocks both sets of HA-specific effector T cells. This observation, together with related findings from other laboratories, is considered to support the idea that T cell recognition is directed against some association of viral and H-2 glycoproteins, as proposed in the original formulation of the "altered self" concept.     

Specificity or affinity of cytotoxic T cells for self H-2K determinants apparently does not change between primary and secondary responses to ectromelia virus infection.

Lysis of virus-infected target cells by virus-specific cytotoxic T cells occurs where donors of T cells and targets share either H-2K or H-2D genes. The effect of four H-2K mutations on virus-induced antigens recognized by cytotoxic T cells from in vitro secondary response to infection was studied. B10.A(5R) cytotoxic T cells (which share the K end of H-2 with the mutant strains, except for the mutated gene(s)) efficiently killed virus-infected macrophage targets from mutant strains B6-H-2bg1 and B6-H-2bg2, were less effective against B6-H-2bh and did not appear to be cytotoxic for B6.C-H-2ba target cells. Conversely, B6-H-Ibg1 and B6-H-2bg2 cytotoxic T cells were more effective in killing virus-infected B10.A(5R) macrophages than B6-H-2bh and B6.C-H-2ba cytotoxic cells respectively. In addition, B6-H-2bg1 and B6-H-2bg2 cells appeared to be only slightly different from wild-type with respect to the interaction between virus-infected cells and T cells. The data obtained suggested that virus-induced antigenic patterns on infected B6.C-H-2ba (mutant) cells are more different antigenically from those on wild-type cells than are those on infected cells from the other mutants, B6-H-2bh, B6-H-2bg1 and B6-H-2bh2. This agrees with previous data using primary cytotoxic T cells and thus suggests that no detectable change in the affinity or specificity of cytotoxic T cell receptors occurs between primary and secondary responses to infection. These findings are also discussed in relation to the exclusion of T cells with receptors for H-2K determinants that are common to the mutants and wild-type, from the response to virus-infected self cells.     

Regulation of primary cytotoxic T lymphocyte responses generated during mixed leukocyte culture with H-2d identical Qa-1-disparate cells

Cytotoxic lymphocyte (CTL) responses are not usually generated during primary mixed leukocyte culture (MLC) with H-2 identical cells. Thus NZB mice are unusual in that their spleen cells do mount CTL responses during primary MLC with H-2d identical stimulator cells; the predominant target antigen for these NZB responses is Qa-1b. Considering the numerous immunoregulatory defects in NZB mice, we postulated that these NZB anti-Qa-1 primary CTL responses were due to an abnormality in T suppressor cell activity. Cellular interactions capable of suppressing NZB anti-Qa-1 primary CTL responses were investigated by using one-way and two-way MLC with spleen cells from NZB mice and other H-2d strains. Although H-2d identical one-way MLC with the use of NZB responders resulted in substantial CTL responses, only minimal CTL responses were detected from two-way MLC with the use of NZB spleen cells plus nonirradiated spleen cells from other H-2d mice. Thus the presence of non-NZB spleen cells in the two-way H-2d identical MLC prevented the generation of NZB CTL. Noncytotoxic mechanisms were implicated in the suppression of the NZB CTL responses during two-way MLC, because only minimal CTL activity was generated when NZB spleen cells were cultured with semiallogeneic, H-2d identical (e.g., NZB X BALB) F1 spleen cells. The observed suppression could be abrogated with as little as 100 rad gamma-irradiation to the non-NZB spleen cells. The phenotype of these highly radiosensitive spleen cells was Thy-1+, Lyt-1+, Lyt-2-, L3T4+. The functional presence of these cells in the spleens of semiallogeneic, H-2d identical F1 mice indicated that their deficiency in NZB mice was a recessive trait. These data suggest that NZB mice lack an L3T4+ cell present in the spleens of normal mice that is capable of suppressing primary anti-Qa-1 CTL responses. This model system should facilitate additional investigations of the cellular interactions and immunoregulatory mechanisms responsible for controlling primary CTL responses against non-H-2K/D class I alloantigens. The model may also provide insight into the immunoregulatory defects of autoimmune NZB mice.     

Genetic analysis of the non-H-2-linked Ir genes controlling the cytotoxic T-cell response to H-Y in H-2 d mice

The T-cell mediated immune responses to the male specific minor histocompatibility antigen H-Y in mice have been studied extensively as a model for immune responses to other weak antigens like tumor antigens or autoantigens. In a recent analysis of the strain distribution of the cytotoxic T-cell (Tc-cell) responsiveness to H-Y, it has been found that genes both within and outside the H-2 complex exert an interactive control. Whereas the H-2 ^b strains all are high responders, independent of their non-H-2 background, other H-2 haplotypes (d, k, and s) only allow for a response if they are combined with certain non-H-2 genes. The H-2-linked immune response genes (Ir-genes) have been previously mapped to the I and K or D region of the H-2 complex, but the mapping of the non-H-2 genes has not yet been established. In this study evidence is presented, using recombinant inbred strains and immunoglobulin heavy chain (Igh) congenic strains of mice, to show that there is more than one non-H-2 Ir-gene involved, that the main controlling genes are not linked to the Igh complex, and that at least one non-H-2 Ir-gene is linked to the H-3 region on chromosome 2. This region includes genes for beta-2-microglobulin (2m), the Ly-mllalloantigen a polymorphic cell surface glycoprotein (Pgp-1), a B-cell specific antigen Ly-4, a transplantation antigen H-3, and genes (Ir-2) controlling the immune response to Ea-1 and H-13.     

Genetic control of immune response to sperm whale myoglobin in mice. I. T lymphocyte proliferative response under H-2-linked Ir gene control.

Studies on the genetic control of immune response to sperm whale myoglobin were initiated. As demonstrated in this paper, the T lymphocyte proliferative response to whale myoglobin is under H-2-linked Ir gene control. Mice of H-2d, H-2f, and H-2s haplotypes were high responders to the myoglobin, whereas haplotypes H-2b, H-2k, H-2p, H-2q, and H-2r were low responders. The Ir gene(s) was localized between H-2K and H2D regions, since the recombinant strain A.TL (KsIkSkDd) was a low responder and A.TH (KsIsSsDd) was a high responder. Further studies with recombinant strains revealed that the expression of the high-responder I-Ad or Ias alleles was sufficient to give a good response, since strains D2.GD (d d b b b b b b) and B10.HTT (s s s s k k k d) were high responders. The expression of the I-Cd allele in strains B10.A (k k k k k d d d) and B10.A(5R) (b b b k k d d d) also gave high response, and thus suggested a second Ir gene, derived from the H-2d haplotype. The finding that expression of the I-Cs allele in B10.S(8R) (k k ? ? s s s s) did not result in high response suggests the lack of the second Ir gene in the high-responder H-2s haplotype.     

Co-dominant restriction by a mixed-haplotype I-A molecule (alpha k beta b) for the lysozyme peptide 52-61 in H-2k x H-2b F1 mice

Helper (CD4+) T lymphocytes recognize protein Ag as peptides associated to MHC class II molecules. The polymorphism of class II alpha- and beta-chains has a major influence on the nature of the peptides presented to CD4+ T lymphocytes. For instance, T cell responses in H-2k and H-2b mice are directed at different epitopes of the hen egg lysozyme (HEL) molecule. The current studies were undertaken with the aim of defining the role of mixed haplotype I-A (alpha k beta b and alpha b beta k) molecules in T cell responses to HEL in (H-2k x H-2b)F1 mice, as well as the nature of the immunogenic peptides of HEL recognized in the context of I-A alpha k beta b and I-A alpha b beta k. A series of HEL-reactive T cell lines and hybridomas derived from MHC class II heterozygous (C57BL/6 x C3H F1) mice were established. Their responsiveness to HEL and synthetic HEL peptides was analyzed with the use of L cells transfected with either I-A alpha k beta b or I-A alpha b beta k as APC. Out of 28 clonal T cell hybridomas tested, 13 (46%) only responded to HEL presented by I-A alpha k beta b, 11 (40%) by I-A alpha b beta k (and to a minor extent I-A alpha k beta k), only 4 (14%) were primarily restricted by I-Ak, and none by I-Ab. All the I-A alpha k beta b-restricted T cell hybridomas responded to the HEL peptide 46-61 and to its shorter fragment 52-61, even at concentrations as low as 0.3 nM. As this determinant has been previously defined as immunodominant for I-Ak but not for I-Ab mice, these results suggest a role for the I-A alpha k chain in the selection and immunodominance of HEL 52-61 in H-2k mice. The fine specificity of I-A alpha k beta b-restricted T cell hybridomas for a series of different HEL peptides around the sequence 52 to 61 suggests that peptide 52-61 binds to I-A alpha k beta b with higher affinity than to I-A alpha k beta k. The peptides recognized in the context of I-A alpha b beta k and I-A alpha k beta k were not identified.     

Control of acute infection with lymphocytic choriomeningitis virus in mice that cannot present an immunodominant viral cytotoxic T lymphocyte epitope.

For controlling infection of the mouse with the lymphocytic choriomeningitis (LCM) virus, CD8+ CTL are essential. In the infected BALB/c mouse the arising LCM virus-specific CTL are exclusively restricted by the class I MHC-encoded molecule L; K- or D-restricted antiviral CTL cannot be detected. Thus, the infected L-deficient BALB/c mutant C-H-2dm2 should not be capable of eliminating the virus. The experimental evidence proves the contrary, which is explained by K- and D-restricted CTL that this mouse generates. Why such cells remain undetectable in BALB/c mice is currently unexplained, because there is no lack of precursors and the corresponding virus Ag is presented. Despite the absence of lytic activity in vitro, other than the one associated with L, transfusion of day 8-immune spleen cells from BALB/c into infected C-H-2dm2 (L-deficient) mice results in accelerated virus elimination from the organs of the latter, which was manifest as soon as 8 h after cell transfer. Furthermore, lytic activity did not attain measurable levels in the recipients' spleens. Obviously, this infection can be terminated by CD8+ T lymphocytes even when these cells' lytic activity is below detectability.     

Recognition of adenovirus E1A gene products on immortalized cell surfaces by cytotoxic T lymphocytes.

The experiments described in this report were designed to examine whether target cells transfected with the adenovirus E1A gene and exhibiting increased susceptibility to lysis by natural killer cells and activated macrophages (J. L. Cook, T. A. Walker, A. M. Lewis, Jr., H. E. Ruley, F. L. Graham, and S. H. Pilder, Proc. Natl. Acad. Sci. USA 83:6965-6969, 1986) also express E1A proteins on their surfaces. MT1A, 12S, and 13S are strain Fischer baby rat kidney (BRK) cell lines immortalized by transfection with plasmids containing only the E1A gene of nononcogenic adenovirus. All of these cell lines were effective in stimulating the generation of cytotoxic T lymphocytes (CTL) in vitro, provided that the cultures were supplemented with an exogenous source of lymphokine and that the responding lymphocytes were from syngeneic Fischer rats previously immunized with a cell line containing the intact E1A gene. HrA2, a Fischer BRK cell line immortalized by transfection with a plasmid containing only exon 1 of the E1A gene, did not generate, nor was it lysed by, E1A-specific CTL. The cytolytic activity of E1A-specific CTL was blocked by antiserum from Fischer rats immunized with purified E1A proteins synthesized in Escherichia coli, supporting the conclusion that an epitope on E1A proteins encoded by the intact E1A gene constitutes part of the CTL target structure on adenovirus-transformed cells. These data suggest that in addition to their functions within host cells, E1A gene products are important immunogenic determinants on the surfaces of adenovirus-transformed cells.     

An immunodominant domain in adenovirus type 2 early region 1A proteins.

Six independent rat hybridoma cell lines producing monoclonal antibodies to human subgroup C adenovirus early region 1A (E1A) proteins were isolated. Competition binding experiments revealed that each of the monoclonal antibodies was directed against the same epitope or overlapping cluster of epitopes on the E1A proteins. Viral E1A deletion mutants and deleted forms of E1A proteins expressed in Escherichia coli were used to localize the antibody recognition sites to sequences between amino acids 23 and 120, encoded within the first exon of the E1A gene. Similarly, polyclonal antisera raised against the trpE-E1A fusion protein, as well as against the native, biologically active E1A protein, were also directed primarily against this immunodominant region.