We have studied the role of histidine 95 (H95) on the pH gating of the cardiac gap junction protein connexin43 (Cx43). Wild through 6 for reviews on the subject). Recently, Spray and Burt2 proposed that acidification-induced uncoupling may result from the protonation of titratable histidine residues present in the cytoplasmic loop (CL) of the cardiac gap junction protein connexin43 (Cx43). Accordingly, the presence of positive charges at positions 95, 126, and/or 142 of Cx43 (positions at which histidine residues reside) should lead to channel closure, whereas the absence of positive charges in those positions may prevent acidification-induced uncoupling. This hypothesis was based on the fact that the imidazole ring of the histidine residue is the only amino acid side chain whose apparent dissociation constant from protons (pKa) falls within the physiological range,7 and it assumes that pH gating would result, at least in part, from a direct effect of the intracellular protons on the conformation of the connexin protein. A role of histidine residues on pH-mediated regulation of protein Received August 9, 1993; accepted February 24, 1994 reduced susceptibility to acidification-induced uncoupling, whereas lysine (a basic residue) was more susceptible to uncoupling than the wild-type protein. The susceptibility to acidification-induced uncoupling was enhanced for the H95Q-A94H mutant when compared with the wild-type mutant, but it was significantly reduced when histidine was placed at position 93 (H95Q-L93H). Our data indicate that a properly placed histidine residue is an important structural element for functional expression as well as for pH regulation of Cx43. The results suggest that the importance of H95 on pH gating may be associated with a possible protonation of this residue on acidification of the intracellular environment. The data are compatible with a recently proposed mechanism modeled after the "ball-and-chain" hypothesis explaining pH gating of Cx43. (Circ Res. 1994;74:1058-1064 Key Words * connexin43 * cardiac gap junctions K pHi * histidine * site-directed mutagenesis function has also been suggested for other membrane proteins (eg, References 8 through 12). More recently, Liu et al13 studied the pH sensitivity of Cx43 when expressed in Xenopus laevis oocytes and demonstrated that deletion of the last 125 amino acids from the carboxyl terminal of Cx43 causes a significant loss in pH, sensitivity, displacing pKa (measured as the value of pHi that causes 50% decrease from maximal junctional conductance [Gj]) from 6.6 in Cx43 channels to 6.1 in the mutant lacking the carboxyl tail. These results demonstrated that the carboxyl end is a necessary domain for normal pH gating of Cx43, whereas residues in the CL cannot by themselves modulate acidification-induced uncoupling. These observations, however, did not imply that the carboxyl end is the only region of the protein involved in pH gating. In fact, it is possible that the carboxyl terminal works in concert with other regions of the protein t...

Sera from approximately 50% of patients with large granular lymphocyte (LGL) leukaemia react with a recombinant human T-cell leukaemia/lymphoma virus (HTLV) transmembrane envelope protein, p21e. Two immunodominant epitopes within env p21e have been defined by reactivity against recombinant proteins GD21 and BA21. In this study sera from 41 patients with LGL leukaemia were examined for reactivity against these recombinant HTLV env proteins. Overall, 21/41 (51%) sera reacted to p21e. Only two sera reacted to GD21. The predominant immunoreactivity against p21e was directed against the BA21 epitope, with 19/41 (46%) sera being BA21 positive. Seroconversion to BA21 protein was also documented. PCR analyses confirmed the low incidence of protypical HTLV sequences (2/41, 5%). These data document an association between BA21 seroreactivity and LGL leukaemia. This finding raises the possibility that such BA21 seroreactivity could be due to cross-reactivity to a cellular or retroviral antigen sharing some amino acid homology with the transmembrane glycoprotein of HTLV.

Summary. The T-cell type of large granular lymphocyte (LGL) leukaemia is a lymphoproliferative disorder characterized by clonal proliferation of CD3 þ LGL, which is often associated with autoimmune disorders. Phenotypic and functional data suggest that leukaemic CD3þ LGL represent activated cytotoxic T lymphocytes (CTL). One mechanism whereby CTL mediate target cell killing is through the Fas/ Fas ligand apoptotic pathway. Fas ligand is expressed by CTL only after activation. In this study we examined seven patients with LGL leukaemia for expression of Fas ligand gene transcripts using reverse transcriptase-polymerase chain reaction (RT-PCR) analyses. We found constitutive expression of Fas ligand gene transcripts in each of the seven patients. Similar up-regulation of Fas ligand gene expression has been observed in mice with autoimmune lymphoproliferative syndromes caused by Fas mutations. However, sequence analyses of the death domain of the Fas gene in LGL leukaemia patients revealed no evidence for mutations. Our findings provide further support for the hypothesis that leukaemic LGL are CTL activated by chronic antigenic stimulation. Constitutive expression of Fas ligand may contribute to the pathogenesis of the neutropenia observed in LGL leukaemia.

The carboxyl terminal cytoplasmic domain of distinct gap junction proteins may play an important role in assembly of functional channels as well as differential responsiveness to pH, voltage, and intracellular second messengers. Oligonucleotide-directed site-specific mutagenesis in a paired Xenopus laevis oocyte expression system was used to examine the expression of mRNAs encoding wild-type and carboxyl terminal mutant connexin43 (Cx43) proteins. Oocytes were stripped, injected with mRNA or distilled water (dH2O), preincubated for 16-20 hours, and then paired for 5-10 hours; this process was followed by electrophysiological recording using the dual voltage-clamp technique. Initial experiments compared the relative junctional conductances (Gjs) in oocyte pairs expressing Cx43 (382 amino acid residues) and two truncated mutants lacking most or a portion of the cytoplasmic carboxyl terminal. The shortest mutant (M241) contained 240 amino acid residues and was devoid of all phosphorylatable serine residues in the cytoplasmic tail; its length approximated the length of liver connexin26. The longest mutant (M257) tested contained 256 amino acid residues, including two serine residues. Oocyte pairs expressing M241 yielded a Gj similar to that of oocytes injected with dH2O, whereas M257 yielded a Gj similar to that of oocytes injected with Cx43. Immunoprecipitation studies showed that Cx43, M257, and M241 proteins were readily detectable in oocytes injected with their respective mRNAs, indicating that the lack of Gj observed with the M241 mRNA was not due to reduced translation. Immunocytochemical studies revealed that wild-type and both truncated mutants were localized to the area of cell-to-cell contact between the paired oocytes, indicating that protein targeting to the membrane was not inhibited in oocytes injected with M241 mRNA. Oocyte pairs expressing mutants in which serine residues were replaced with nonphosphorylatable amino acids (serine codon No. 255 AGC was converted to GCC, alanine, designated as M255S----A, and serine codon No. 244 AGC was converted to GGC, glycine, designated as M244S----G) showed Gjs similar to M257, indicating that these serine residues and, by inference, their phosphorylation state are not critical for expression of functional channels. The importance of the length of the carboxyl terminus was assessed by comparing the Gjs in a series of mutants that were intermediate in length between M257 and M241. Gradual shortening of the carboxyl terminus produced a gradual reduction of Gj relative to M257. However, simple deletion of amino acid residues 241-257 from the wild-type Cx43 did not affect Gj relative to M257.(ABSTRACT TRUNCATED AT 400 WORDS)

Samples were obtained from 53 large granular lymphocytic leukemia (LGLL) patients and 10,000 volunteer blood donors (VBD). Sera were screened in an HTLV-1 enzyme immunoassay (EIA) and further analyzed in peptide-specific Western blots (WB). DNAs were analyzed by HTLV-1, -2, -3, and -4-specific PCR. Forty four percent of LGLL patients vs. 0.12 % of VBD had anti-HTLV antibodies via EIA (p < 0.001). WB and PCR revealed that four LGLL patients (7.5%) vs. one VBD patient (0.01%) were infected with HTLV-2 (p < 0.001), suggesting an HTLV-2 etiology in a minority of cases. No LGLL patient was positive for HTLV-1, -3, or -4, whereas only one EIA-positive VBD was positive for HTLV-1 and none for HTLV-3 or -4. The HTLV EIA-positive, PCR-negative LGLL patients' sera reacted to epitopes within HTLV p24 gag and gp21 env. Other then the PTLV/BLV viruses, human endogenous retroviral element HERV K10 was the only sequence homologous to these two HTLV peptides, raising the possibility of cross-reactivity. Although three LGLL patients (5.7%) vs. none of 110 VBD patients tested positive for antibodies to the homologous HERV K10 peptide (p = 0.03), the significance of the anti-HTLV seroreactivity observed in many LGLL patients remains unclear. Interestingly, out of 36 HTLV-1-positive control subjects, 3 (8%) (p = 0.014) were positive for antibodies to HERV K10; all three had myelopathy. Out of 64 HTLV-2-positive control subjects 16 (25%) (p = <0.001) were positive for HERV K10 antibodies, and 4 (6%) of these had myelopathy. Out of 22 subjects with either HTLV-1 or -2 myelopathy, 7 (31.8%) were positive for HERV K10 antibodies, and out of 72 HTLV-infected subjects without myelopathy, 12 (16.7%) were positive for anti-HERV K10 antibodies (p = 0.11). The prevalence of anti-HERV K10 antibodies in these populations and the clinical implications thereof need to be pursued further.