Veratridine-induced high-frequency asynchronous release of inhibitory transmitter quanta in crayfish nerve-muscle synapses superfused with normal and low-calcium saline

Crayfish fibres of opener muscles were voltage clamped toE=–80 mV membrane potential (T=19–22°C), and veratridine (10–100 mol/l) was added to the superfusate. Within 30–60 s this caused large fluctuations of the clamp current due to vigorous asynchronous quantal release from the inhibitory nerve terminals along the muscle fibre. Excitatory postsynaptic receptors were previously desensitized by application of 5 mmol/l glutamate. Current fluctuations were evaluated by means of the noise analysis technique. Typically, 100 mol/l veratridine increased instantaneously the quantal release rate ñ from ñ<1 quantum/s toñ10,000 quanta/s. Thereafter, ñ declined exponentially with a time constant of 70s. On average, about 500,000 inhibitory quanta could be liberated in this way from the terminals on a single muscle fibre of 1 mm length. Serotonin (1 mol/l) facilitated the effect of lower veratridine concentrations (1–10 mol/l). In opener muscles veratridine-induced asynchronous quantal release showed little dependence on the bath concentration of Ca²⁺. The opposite was found for fibres of the superficial abdominal extensor muscle. Beside postsynaptic current fluctuations, veratridine elicited slowly changing average postsynaptic DC-currents which could be explained partly by superposition of individual inhibitory quantal currents. These DC-currents suggest that beside inhibitory quantal release another factor activates inhibitory postsynaptic receptors after application of veratridine.This investigation was supported by the Deutsche Forschungsgemeinschaft, SFB 220     

Noise analysis of acetylcholine induced current at neuromuscular junctions of the mouse diaphragm

Miniature end-plate currents (mepcs) and membrane noise elicited by acetylcholine (ACh) iontophoresis were investigated at neuromuscular junctions of the mouse diaphragm. All the experiments were performed at a holding potential of -70 mV at a temperature of 19 degrees C. The equilibrium potential of the ACh response was estimated to be near zero; the mepcs displayed a peak amplitude of 2.46 +/- 0.13 nA (mean +/- S.E.) and relaxed exponentially with a time constant to 1.63 +/- 0.11 msec. Single ACh-activated channels had a conductance of 26.5 +/- 1.5 pS and a mean life time of 1.69 +/0- 0.13 msec.     

Zinc competitively inhibits calcium-dependent release of transmitter at the mouse neuromuscular junction

The effect of zinc on the release of transmitter was investigated in preparations of mouse diaphragm by conventional microelectrode techniques. The frequency (FP of miniature end-plate potentials (MEPPs), elevated by Ca²⁺ in high K⁺ medium, was reduced by zinc in a concentration-dependent fashion. When the extracellular concentration of Ca²⁺ ([Ca²⁺]_o) was varied in the absence of zinc, a linear relationship between log(F) and log([Ca²⁺]_o) was obtained. When the effect of zinc was depicted graphically, it was found that zinc shifted the relationship between log(F) and log([Ca²⁺]_o) to the right, with respect to the control in the absence of zinc, without altering the slope. Zinc also reduced the quantal content (m) of end-plate potentials (EPPs). As [Ca²⁺]_o was varied in the absence of zinc, a linear relationship between ln(m) and ln([Ca²⁺]_o) was observed. Zinc shifted this linear relationship between ln(m) and ln([Ca²⁺]_o) to the right, with respect to the control, without altering the slope. Thus, zinc reduced both the asynchronous and the phasic release of transmitter. These results suggest that zinc competes with Ca²⁺, and this conclusion is confirmed by examination of a modified Lineweaver-Burk plot of the data. Zinc probably inhibits the entry of Ca²⁺ into the nerve terminals, thereby inhibiting transmitter release.     

Modulation of available vesicles and release kinetics at the inhibitor of the crayfish neuromuscular junction

We have investigated the effect of serotonin (5-HT) and okadaic acid (OA) on presynaptic processes at the crayfish inhibitory neuromuscular junction. Two different physiological parameters of transmitter release were examined: release kinetics and the size of the readily releasable pool of vesicles (RRP). Using a paired pulse stimulus and high frequency trains, we established that a single broad action potential, recorded in 20 mM tetraethylammonium and 1 mM 4-amino-pyridine, released the RRP in its entirety. Thus, by measuring the amplitude of inhibitory postsynaptic potential (IPSC) we were able to directly assess the effects of 5-HT and OA on the RRP. Serotonin at 200 nM and OA at 2.5 μM each significantly increased IPSC above control levels and the effects of these two modulators were comparable. Both modulators also induced a leftward shift in the rising phase of IPSC, i.e. an apparent acceleration in release kinetics. The shift caused by OA was significantly more pronounced than that induced by 5-HT. This apparent acceleration in release was not associated with a corresponding change in the presynaptic Ca²⁺ transient measured at a 2 kHz resolution, suggesting that modulation was not due to an acceleration in Ca²⁺ channel kinetics. In view of the comparable increase in the size of the RRP by the modulators, the differential modulation of release kinetics suggests that these two parameters may be modulated by separate biochemical processes.     

Topographic comparison of neuromuscular junctions in mouse slow and fast twitch muscles

The neuromuscular junctions of mammalian slow and fast twitch muscles are activated differently in vivo and show corresponding physiological differences in vitro, but the structural basis or consequences of these differences are relatively unexplored. Therefore, neuromuscular junctions of mouse fast (extensor digitorum longus) and slow (soleus) twitch muscles were compared by use of new scanning and light microscopy techniques. In both muscles, the endplate appeared as an elliptical area raised to a variable extent above the surrounding sarcolemma and containing the primary clefts. In most soleus endplates, this raised surface area was considerably higher and wider and about three times larger than in extensor digitorum longus. In addition, the primary cleft area was about two-fold greater in soleus than in extensor digitorum longus, even though cleft length was the same. The primary clefts formed either an elliptical shape along the outer margin of the endplate with inward-directed branches or a group of relatively rectilinear dendritic branches orthogonally oriented to one another. The latter type was most frequent in soleus and the elliptical type in extensor digitorum longus. Corresponding patterns of nerve terminal arborizations were seen by light microscopy. Although nerve terminal areas were the same in fast and slow muscles, in the former, numerous diverticulae significantly increased the length of the nerve terminal outline. The possible physiological significance of the different synaptic structure of slow and fast muscle is discussed.     

Effect of purines on calcium-independent acetylcholine release at the mouse neuromuscular junction

At the mouse neuromuscular junction, activation of adenosine A(1) and P2Y receptors inhibits acetylcholine release by an effect on voltage dependent calcium channels related to spontaneous and evoked secretion. However, an effect of purines upon the neurotransmitter-releasing machinery downstream of Ca(2+) influx cannot be ruled out. An excellent tool to study neurotransmitter exocytosis in a Ca(2+)-independent step is the hypertonic response. Intracellular recordings were performed on diaphragm fibers of CF1 mice to determine the action of the specific adenosine A(1) receptor agonist 2-chloro-N(6)-cyclopentyl-adenosine (CCPA) and the P2Y(12-13) agonist 2-methylthio-adenosine 5'-diphosphate (2-MeSADP) on the hypertonic response. Both purines significantly decreased such response (peak and area under the curve), and their effect was prevented by specific antagonists of A(1) and P2Y(12-13) receptors, 8-cyclopentyl-1,3-dipropylxanthine (DPCPX) and N-[2-(methylthioethyl)]-2-[3,3,3-trifluoropropyl]thio-5'-adenylic acid, monoanhydride with dichloromethylenebiphosphonic acid, tetrasodium salt (AR-C69931MX), respectively. Moreover, incubation of preparations only with the antagonists induced a higher response compared with controls, suggesting that endogenous ATP/ADP and adenosine are able to modulate the hypertonic response by activating their specific receptors. To search for the intracellular pathways involved in this effect, we studied the action of CCPA and 2-MeSADP in hypertonicity in the presence of inhibitors of several pathways. We found that the effect of CPPA was prevented by the calmodulin antagonist N-(6-aminohexil)-5-chloro-1-naphthalenesulfonamide hydrochloride (W-7) while that of 2-MeSADP was occluded by the protein kinase C antagonist chelerythrine and W-7. On the other hand, the inhibitors of protein kinase A (N-(2[pbromocinnamylamino]-ethyl)-5-isoquinolinesulfonamide, H-89) and phosphoinositide-3 kinase (PI3K) (2-(4-morpholinyl)-8-phenyl-4H-1-benzopyran-4-one hydrochloride, LY-294002) did not modify the modulatory action in hypertonicity of both purines. Our results provide evidence that activation of A(1) and P2Y(12-13) receptors by CCPA and 2-MeSADP inhibits ACh release from mammalian motor nerve terminals through an effect on a Ca(2+)-independent step in the cascade of the exocytotic process. Since presynaptic calcium channels are intimately associated with components of the synaptic vesicle docking and fusion processes, further experiments could clarify if the actions of purines on calcium channels and on secretory machinery are related.     

Presynaptic inhibition of spontaneous acetylcholine release mediated by P2Y receptors at the mouse neuromuscular junction

At the neuromuscular junction, ATP is co-released with the neurotransmitter acetylcholine (ACh) and once in the synaptic space, it is degraded to the presynaptically active metabolite adenosine. Intracellular recordings were performed on diaphragm fibers of CF1 mice to determine the action of extracellular ATP (100 muM) and the slowly hydrolysable ATP analog 5'-adenylylimidodiphosphate lithium (betagamma-imido ATP) (30 muM) on miniature end-plate potential (MEPP) frequency. We found that application of ATP and betagamma-imido ATP decreased spontaneous secretion by 45.3% and 55.9% respectively. 8-Cyclopentyl-1,3-dipropylxanthine (DPCPX), a selective A(1) adenosine receptor antagonist and alpha,beta-methylene ADP sodium salt (alphabeta-MeADP), which is an inhibitor of ecto-5'-nucleotidase, did not prevent the inhibitory effect of ATP, demonstrating that the nucleotide is able to modulate spontaneous ACh release through a mechanism independent of the action of adenosine. Blockade of Ca(2+) channels by both, Cd(2+) or the combined application of nitrendipine and omega-conotoxin GVIA (omega-CgTx) (L-type and N-type Ca(2+) channel antagonists, respectively) prevented the effect of betagamma-imido ATP, indicating that the nucleotide modulates Ca(2+) influx through the voltage-dependent Ca(2+) channels related to spontaneous secretion. betagamma-Imido ATP-induced modulation was antagonized by the non-specific P2 receptor antagonist suramin and the P2Y receptor antagonist 1-amino-4-[[4-[[4-chloro-6-[[3(or4)-sulfophenyl] amino]-1,3,5-triazin-2-yl]amino]-3-sulfophenyl] amino]-9,10-dihydro-9,10-dioxo-2-anthracenesulfonic acid (reactive blue-2), but not by pyridoxal phosphate-6-azo(benzene-2,4-disulfonic acid) tetrasodium salt (PPADS), which has a preferential antagonist effect on P2X receptors. Pertussis toxin and N-ethylmaleimide (NEM), which are blockers of G(i/o) proteins, prevented the action of the nucleotide, suggesting that the effect is mediated by P2Y receptors coupled to G(i/o) proteins. The protein kinase C (PKC) antagonist chelerythrine and the calmodulin antagonist N-(6-aminohexil)-5-chloro-1-naphthalenesulfonamide hydrochloride (W-7) occluded the effect of betagamma-imido ATP, while the protein kinase A (PKA) antagonist KT-5720 and the inhibitor of the calcium/calmodulin-dependent protein kinase II (CAMKII) KN-62 failed to do so. betagamma-Imido ATP did not affect 10, 15 and 20 mM K(+)-evoked release and application of reactive blue-2 before incubation in high K(+) induced a higher asynchronous secretion. Thus, our results show that at mammalian neuromuscular junctions, ATP induces presynaptic inhibition of spontaneous ACh release due to the modulation of Ca(2+) channels related to tonic secretion through the activation of P2Y receptors coupled to G(i/o) proteins. We also demonstrated that at increasing degrees of membrane depolarization evoked by K(+), endogenously released ATP induces presynaptic inhibition as a means of preventing excessive neurotransmitter secretion.     

Effects ofAnemonia sulcata toxin II on presynaptic currents and evoked transmitter release at neuromuscular junctions of the mouse

The effect ofAnemonia sulcata toxin II (ATX-II) on the amount of transmitter released by nerve impulses was investigated in motor end-plates of the mouse. ATX-II (80 nM) caused repetitive end-plate potentials in response to a single nerve stimulus and a 3- to 4-fold increase in the quantal content of the phasic end-plate potential. This increase is less than what would be expected if ATX-II induced plateau action potentials at the motor endings. To solve this discrepancy presynaptic currents were recorded by focal extracellular electrodes. It was found that the K current present at the endings is strong enough to prevent the development of presynaptic plateau action potentials, in contrast to what has been observed in other excitable membranes (unmyelinated axons, nodes of Ranvier and skeletal muscle fibres). By using tetraethylammonium and 3,4-diaminopyridine to block K channels and Co²⁺ to block Ca channels, ATX-II allowed the development of prolonged plateau responses at the endings upon motor nerve stimulation. These results suggest that the mouse motor endings are endowed with a relatively powerful K channel system, which effectively controls the amount of presynaptic depolarization.     

Acetylcholine receptors and nerve terminal distribution at the neuromuscular junction of non‐obese diabetic mice

Skeletal muscle is one of the main targets of the metabolic alterations in diabetes, in which protein synthesis is markedly reduced followed by increased proteolysis. Ultrastructural and functional changes in the presynaptic compartment of the neuromuscular junction (NMJ) have been demonstrated, but little attention has been paid to the proteins in the postsynaptic muscle fiber membrane. In the present work, we studied the changes in acetylcholine receptors (AChRs) and nerve terminal distribution in the NMJ of non‐obese diabetic (NOD) mice. The sternomastoid muscles of adult female NOD mice were double‐labeled for AChR and nerve terminal observation by fluorescence and reflected light confocal microscopy. In 62.4% of the diabetic endplates, AChR branches broke apart into receptor islands that stained less than in the normal mice. These patches had regular junctional folds. At most of the endplates studied, the nerve terminals colocalized with AChRs, and sprouts were seen in 10% of the diabetic endplates. The intramuscular nerve branches and axons in the nerve to the sternomastoid muscle showed no degenerative disorders. These results suggest that metabolic alterations in the diabetic muscle fiber can affect the distribution and expression of molecules, such as AChRs, in the postsynaptic membrane of the neuromuscular junction. Anat Rec 267:112–119, 2002. © 2002 Wiley‐Liss, Inc.     

The role of extracellular calcium in exo- and endocytosis of synaptic vesicles at the frog motor nerve terminals

In the present study we combined FM 1-43 imaging and electrophysiological recording of miniature end-plate currents (MEPCs) to determine the role of extracellular calcium in synaptic vesicle exo- and endocytosis at the frog motor nerve terminals. We replaced extracellular Ca2+ ions with other bivalent cations (Sr2+, Ba2+, Cd2+, Mg2+) or used a calcium-free solution and monitored fluorescent staining of the nerve terminals in the presence of caffeine, which promotes the release of Ca2+ from intracellular stores. Caffeine has induced FM1-43 internalization only in the presence of bivalent cations in the external solution. The exposure of the neuromuscular junction to caffeine in a calcium-free solution caused a reversible failure of FM 1-43 loading and an increase in the nerve terminal width. This effect of a calcium-free solution was not due to a decrease in exocytosis, because caffeine-induced FM1-43 unloading from the previously loaded nerve terminals, as well as a degree of the MEPCs frequency increase, was unchanged. We conclude that the presence of Ca2+ or other bivalent cations in extracellular space is necessary for endocytosis but not for exocytosis of synaptic vesicles, while transmitter release is promoted by efflux of Ca2+ from intracellular stores. The effect of extracellular Ca2+ on endocytosis might be driven by the non-specific interactions with membrane lipids.     

Effects of dopamine and dibutyryl cyclic adenosine monophosphate on delayed release of transmitter at the rat neuromuscular junction

The effects of dopamine and dibutyryl cyclic adenosine monophosphate (db-cAMP) on delayed release of transnutter were studied in vitro in the phrenic nervediaphragm muscle preparation of the rat using intracellular recording techniques. Dopamine at 1 × 10^–4 mol 1^–1 prevented the initial facilitation of delayed release of transmitter. This inhibitory phase was transformed into a transient facilitation of delayed release. We observed that dopamine hyperpolarized muscle fibres by about 9%. Thus motor nerve terminals may also have been hyperpolarized by dopamine; however, it is unlikely that this hyperpolarization explains the observed effects on delayed release of transmitter. Db-cAMP at 1 × 10^–3 mol 1^–1 predominantly augmented delayed release of acetylcholine. These effects of dopamine and db-cAMP on delayed release of transmitter are discussed in terms of a modulation of calcium fluxes in the presynaptic nerve terminal.     

The fibrillar structure of cementum and dentin at the cemento-dentinal junction in rat molars

The cemento-dentinal junction was examined in demineralized rat molars with complete roots by scanning electron microscopy combined with NaOH maceration. It is established that the NaOH maceration removes interfibrillar substances and cells from connective tissues selectively without structural damage to collagen fibrils. The cementum was detached from the dentin by the maceration. The inner cementum surface facing the dentin and the outer dentin surface facing the cementum were observed. In acellular cementum, both the outer dentin surface and the inner cementum surface had a smooth appearance. There was little indication of fibrils intermingling between dentin and cementum. In contrast, both the inner cementum surface and outer dentin surface in cellular cementum had an uneven appearance due to the irregular arrangement of collagen fibrils. Point-like protrusions of fibril bundles were observed on both surfaces. Some (not all) of these point-like protrusions appeared to correspond to places of fibrillar intermingling between dentin and cementum.     

Characterization of acetylcholine release and the compensatory contribution of non-Ca(v)2.1 channels at motor nerve terminals of leaner Ca(v)2.1-mutant mice

The severely ataxic and epileptic mouse leaner (Ln) carries a natural splice site mutation in Cacna1a, leading to a C-terminal truncation of the encoded Ca(v)2.1 alpha(1) protein. Ca(v)2.1 is a neuronal Ca(2+) channel, mediating neurotransmitter release at many central synapses and the peripheral neuromuscular junction (NMJ). With electrophysiological analyses we demonstrate severely reduced ( approximately 50%) neurotransmitter release at Ln NMJs. This equals the reduction at NMJs of Cacna1a null-mutant (Ca(v)2.1-KO) mice, which display a neurological phenotype remarkably similar to that of Ln mice. However, using selective Ca(v) channel blocking compounds we revealed a compensatory contribution profile of non-Ca(v)2.1 type channels at Ln NMJs that differs completely from that at Ca(v)2.1-KO NMJs. Our data indicate that the residual function and presence of Ln-mutated Ca(v)2.1 channels precludes presynaptic compensatory recruitment of Ca(v)1 and Ca(v)2.2 channels, and hampers that of Ca(v)2.3 channels. This is the first report directly showing at single synapses the deficits and plasticity in transmitter release resulting from the Ln mutation of Cacna1a.     

Identification of vascular sphincters at the junction between alveolar capillaries and pulmonary venules of the mouse lung

Background: A peculiar feature of lung circulation in the lung is the pronounced variations in blood volume observed in alveolar capillaries that occur because of the changes in the conformation of the alveolar wall that are associated with the respiratory movements. This phenomenon has led to the postulate that mechanisms of postcapillary control of blood flow are to be present in the lung vessels. In the present study we searched for microanatomical evidence of vascular sphincters in the deep lung tissue of mice, namely in alveolar capillaries and pulmonary veins. Methods: We have used scanning electron microscopy (SEM) to examine two types of samples of normal lung tissue of CD-1 mice: 1) vascular corrosion casts made by vascular perfusion with Mercox® resin, and 2) routinely made gold/platinum-coated replicas of sectioned lung tissue. Results: Careful scrutiny of the vessels of the deep lung tissue led to the identification of sphincters in alveolar capillaries. These sphincters were located at the junction between capillary and pulmonary veins. They corresponded to areas to the vascular wall showing circular swellings where a radial organization was observed, since they were made up of alternating grooves and bulges. Transmission electron microscopy showed that smooth muscle cells participated in the formation of the sphincters. Conclusions: Our data reveal a new location for vascular sphincters in pulmonary vessels and, because these novel sphincters are located at the capillary-vein junction, they offer a structural setting for the existence of postcapillary control of blood flow in the pulmonary circulation of mice. © 1995 Wiley-Liss, Inc.     

Non-quantal release of transmitter at mouse neuromuscular junction and its dependence on the activity of Na+−K+ ATP-ase

Summary The effect of val⁵-angiotensin II on steady-state sodium concentration gradients (c_Na) was studied in rat proximal tubules by stationary micro-perfusion combined with perfusion of the peritubular capillaries. Angiotensin added to the peritubular perfusion fluid had a biphasic action with stimulation of sodium reabsorption at low doses (10^–12–10^–10 M) and inhibition at high doses (3×10^–7–3×10^–6 M). Stimulation of transport was also observed with intraluminal angiotensin but only at a dose of 10^–9M. Transepithelial potential difference was calculated from the steady-state chloride distribution; no significant change was observed at low (10^–11M) or high (10^–6M) concentrations and a direct action on sodium transport is postulated. This biphasic effect is discussed in relation to the responses of the intact kidney to intra-renal infusion of angiotensin, and to the control of tubulo-glomerular feed-back.     

N type Ca2+ channels and RIM scaffold protein covary at the presynaptic transmitter release face but are components of independent protein complexes

Fast neurotransmitter release at presynaptic terminals occurs at specialized transmitter release sites where docked secretory vesicles are triggered to fuse with the membrane by the influx of Ca²⁺ ions that enter through local N type (CaV2.2) calcium channels. Thus, neurosecretion involves two key processes: the docking of vesicles at the transmitter release site, a process that involves the scaffold protein RIM (Rab3A interacting molecule) and its binding partner Munc-13, and the subsequent gating of vesicle fusion by activation of the Ca²⁺ channels. It is not known, however, whether the vesicle fusion complex with its attached Ca²⁺ channels and the vesicle docking complex are parts of a single multifunctional entity. The Ca²⁺ channel itself and RIM were used as markers for these two elements to address this question. We carried out immunostaining at the giant calyx-type synapse of the chick ciliary ganglion to localize the proteins at a native, undisturbed presynaptic nerve terminal. Quantitative immunostaining (intensity correlation analysis/intensity correlation quotient method) was used to test the relationship between these two proteins at the nerve terminal transmitter release face. The staining intensities for CaV2.2 and RIM covary strongly, consistent with the expectation that they are both components of the transmitter release sites. We then used immunoprecipitation to test if these proteins are also parts of a common molecular complex. However, precipitation of CaV2.2 failed to capture either RIM or Munc-13, a RIM binding partner. These findings indicate that although the vesicle fusion and the vesicle docking mechanisms coexist at the transmitter release face they are not parts of a common stable complex.     

丹参注射液对冻融小鼠卵巢移植早期光镜结构及超微结构的影响

目的 探讨丹参注射液对冻融小鼠卵巢同种异体移植早期光镜结构及超微结构的影响.方法 收集1日龄小鼠卵巢,慢冻速融后移植至F1代8～12周雄鼠的肾被膜下,分别于移植后第2天(2d)、第7天(7d)和第14天(14d)回收两组卵巢移植物.光镜和电镜技术观察卵巢组织冻融前后和移植前后的组织形态学改变、计数卵泡数量.结果 小鼠卵巢经冻融后原始卵泡的存活率为87.9%.光、电镜观察卵巢组织移植后7d卵泡细胞和颗粒细胞结构欠完整,移植后14d卵泡结构恢复正常.移植后14d丹参组各级卵泡数及卵泡存活率(30.1%)多于生理盐水组(13.2%),差异显著(P＜0.05).结论 结合光镜和电镜观察能更全面地评价冻融过程和移植过程对小鼠卵巢组织的影响,移植早期的缺血缺氧对小鼠卵巢组织造成了远比冻融过程更严重的损伤;丹参注射液可能促进移植后冻融小鼠卵泡发育,减少细胞凋亡.     

Complete nucleotide sequence of a Coxsackievirus B-4 strain capable of establishing persistent infection in human pancreatic islet cells: effects on insulin release,proinsulin synthesis,and cell morphology

The aim of the present investigation was to study the effect of infection of human pancreatic islet cells with a strain (VD2921) of Coxsackie B virus serotype 4 capable of establishing persistent infection in these cells, as well as to sequence the strain, to study the determinants of virulence and persistence. Groups of islets were infected and assessments of proinsulin, insulin content, and virus replication were made. Insulin release in response to high glucose was measured. Infected and control islets displayed a strong insulin response to high glucose 3-4 days as well as 7-8 days post-infection (dpi). At 11-17 dpi, the infected islets did not respond at all to high glucose, and the response of the control islets was at this late time point somewhat reduced. The insulin and proinsulin content of the infected islets did not differ significantly from that of the control islets. TCID(50) titrations showed that the VD2921 strain replicated in the islet cells during the whole study. Electron microscopic examination of infected islets did not reveal any virus-induced changes of cell morphology compared with the controls, although higher magnifications of the infected beta-cells showed virus-like particles in the cytoplasm. These results show that certain strains of Coxsackievirus B-4 in vitro can establish a persistent infection that might mimic, the more gradual loss of beta-cell function seen during the clinical course of autoimmune diabetes. The ability of this Coxsackievirus B-4 strain to establish a persistent infection might be due to substitution of 11 amino acids located at the surface of the structural protein VP1, adjacent to the predicted receptor binding canyon of the virus.