Syntaxins 1 through 4 are expressed in the mouse retina
Immunoblotting of retinal and brain membrane homogenates showed that each anti-syntaxin antibody recognized a single protein band of the appropriate size (Fig. ), indicating that each of the four syntaxin isoforms associated with vesicular trafficking to the plasma membrane was present in the mouse retina and brain. Although syntaxins 1 through 4 were all expressed in retina and in brain, relative expression levels differed between brain and retina in an isoform-specific manner. Retina and brain showed roughly similar expression of syntaxin 1, retina expressed more syntaxin 3 than brain, and brain expressed more syntaxin 2 and syntaxin 4 than retina.
Figure 1 Syntaxins 1 through 4 are differentially distributed in the synaptic layers of the retina. A: Syntaxin 1 is localized to amacrine cell bodies (A) and processes in the inner plexiform layer (IPL). B: Syntaxin 2 also is localized to amacrine cell bodies (more ...)
Each syntaxin isoform showed a unique distribution in the retina and was localized primarily to the synaptic layers (Fig. ). Syntaxin 1 labeling was widely distributed in the inner plexiform layer (IPL) and in amacrine cell bodies, but was present only in very few terminals in the outer plexiform layer (OPL). A prominent characteristic of syntaxin 1 labeling was its presence in discrete puncta embedded in a background of diffuse labeling. Syntaxin 2 labeling also was widely distributed in the IPL and in the cell bodies of amacrine cells, but was more punctate in appearance than labeling for syntaxin 1. There was little labeling for syntaxin 2 in the OPL, similar to syntaxin 1. Labeling for syntaxin 3 was present in the cell bodies, inner segments and synaptic terminals of photoreceptors, in bipolar cell bodies and axons, and numerous terminals in the IPL. Syntaxin 4 labeling was the most restricted of all the isoforms and was present in numerous puncta in the OPL, a distinct stratum of puncta at the distal edge of the IPL and in other puncta scattered deeper in the IPL. Syntaxin 4 labeled puncta also were observed in the inner nuclear layer (INL) and some blood vessels (see below).
The isoform-specific differences in expression level and localization suggest that each syntaxin isoform has unique trafficking functions in the retina. To better understand the distribution and potential functional roles of the various syntaxin isoforms in the retina, we performed double-labeling for each syntaxin isoform in combination with a variety of cell and synapse-specific markers with known distribution in the retina.
Syntaxin 1 and 3 are presynaptic and differentially expressed between conventional and ribbon synapses
Syntaxin 1 and 3 isoforms have been examined in retina previously [12
]. Syntaxin 1 has been localized to the conventional synapses of amacrine cells, while syntaxin 3 has been localized to glutamatergic ribbon synapses of photoreceptors and bipolar cells [12
Double labeling for syntaxin 1 and synapsin I, a marker specific for conventional synapses made by amacrine cells [19
], confirmed the localization of syntaxin 1 to amacrine cells and their processes and synapses in the IPL. Labeling for syntaxin 1 and synapsin I colocalized extensively (Fig. ), confirming the presence of syntaxin 1 at conventional amacrine cell synapses. A substantial portion of syntaxin 1 labeling was diffusely distributed in the IPL and did not correspond precisely to synapsin I labeled puncta, suggesting that syntaxin 1 labeling was not restricted only to the synaptic active zone.
Figure 2 Syntaxin 1 is associated with conventional synapses in the inner plexiform layer (IPL). A–C: Confocal microscopy of double labeling for syntaxin 1 and synapsin I, a marker for conventional synapses made by amacrine cells. There is appreciable (more ...)
Previous results indicate that syntaxin 3 is associated with the glutamatergic ribbon synapses of photoreceptors and bipolar cells [12
]. However, it is not known whether syntaxin 3 is present in the synapses of a recently identified "glutamatergic" amacrine cell that expresses vesicular glutamate transporter 3 (VGLUT3) [24
], which would imply a functional association of syntaxin 3 with glutamatergic transmission rather than with ribbon synapses exclusively. To test this, we examined colocalization of syntaxin 3 with several markers specific for conventional synapses, glutamatergic ribbon synapses, and VGLUT3.
Double labeling for syntaxin 3 and syntaxin 1 directly demonstrated that syntaxin 1 and 3 were restricted to different terminals, although there often was close apposition of syntaxin 3- and syntaxin 1-containing terminals (Fig. ). Double labeling for syntaxin 3 and the conventional synapse marker synapsin I produced similar results (not shown) indicating that syntaxin 3 was absent from most conventional synapses. Double labeling for syntaxin 3 and vesicular glutamate transporter 1 (VGLUT1), which is expressed at all photoreceptor and bipolar cell ribbon synapses [27
], confirmed directly that syntaxin 3 was expressed at the glutamatergic ribbon synapses of photoreceptors and bipolar cells (Fig. ). All photoreceptor and bipolar cell terminals showed double labeling for VGLUT1 and syntaxin 3, indicating that syntaxin 3 was, in fact, expressed in all ribbon synaptic terminals. Syntaxin 3 labeling, however, was distributed more broadly than VGLUT1 labeling, suggesting that syntaxin 3 was not strictly confined to the active zone.
Figure 3 Syntaxin 3 does not colocalize with syntaxin 1. A–C: Syntaxin 3 is localized to the glutamatergic photoreceptor terminals in the outer plexiform layer (OPL) and numerous bipolar cell terminals in the inner plexiform layer (IPL). Labeling is also (more ...)
Figure 4 Syntaxin 3 is associated with the glutamatergic ribbon synapses of photoreceptors and bipolar cells. A–C: Labeling for syntaxin 3 and VGLUT1, a marker for ribbon synapses, colocalize extensively in the photoreceptor terminals in the OPL and in (more ...)
The recent discovery of a small population of putative glutamatergic amacrine cells expressing VGLUT3 [24
], raised the issue of whether syntaxin 3 also might be expressed at conventional glutamatergic synapses in the retina in addition to ribbon synapses. Double labeling showed that syntaxin 3 was not expressed at conventional synapses of VGLUT3 amacrine cells, although close apposition of some VGLUT3-positive and syntaxin 3-positive terminals was observed (Fig. ). This result confirmed that syntaxin 3 was associated specifically with the release machinery of ribbon synapses and not with glutamatergic transmission in the retina in general.
Figure 5 Syntaxin 3 is not expressed at putative conventional glutamatergic synapses. A. Syntaxin 3 labeling is present in numerous terminals in the inner plexiform layer (IPL), corresponding to the ribbon synapse-containing terminals of bipolar cells (see text (more ...)
Syntaxin 2 is expressed specifically by amacrine cells
Syntaxin 2 labeling was found in amacrine cell bodies and numerous puncta in the IPL and was grossly similar to the distribution of syntaxin 1 labeling (see Fig. ), suggesting that syntaxin 2 was expressed by amacrine cells. Amacrine cells are typically inhibitory, with about half of the amacrine cell population being GABAergic and about half being glycinergic [15
]. To assess whether syntaxin 2 might be associated specifically with either GABAergic or glycinergic neurotransmission, retinal sections were double labeled for syntaxin 2 and either the 65 kDa isoform of glutamic acid decarboxylase (GAD-65) or glycine transporter 1 (GlyT1) to identify GABAergic and glycinergic amacrine cells, respectively (Fig. ). Syntaxin 2 labeling colocalized extensively with labeling for GAD-65 and GlyT1 in amacrine cell bodies, indicating that both GABAergic and glycinergic amacrine cell types expressed syntaxin 2 (Fig. and ). Within the IPL, where the GABAergic and glycinergic synapses of the amacrine cells reside, there was scant colocalization of labeling for syntaxin 2 and GAD-65, which labels the presynaptic terminals of most GABAergic amacrine cells (Fig. ). Similarly, syntaxin 2 labeling showed only limited colocalization with GlyT1 labeling (Fig. ), although colocalization of syntaxin 2 and GlyT1 labeling was seen more frequently than colocalization of syntaxin 2 and GAD-65.
Figure 6 Syntaxin 2 is expressed in GABAergic and glycinergic amacrine cells. A–C: Labeling for syntaxin 2 colocalizes with labeling for GAD-65 in numerous GABAergic amacrine cell bodies (*) in the inner nuclear layer (INL). However, syntaxin 2 labeling (more ...)
Syntaxin 2 shows little colocalization with syntaxin 1 and does not appear to be presynaptic
The localization of syntaxin 2 labeling to amacrine cells, suggested that syntaxin 2 potentially might have some redundancy with syntaxin 1. This was directly examined by double labeling for syntaxin 2 and syntaxin 1. Some colocalization of syntaxin 1 and syntaxin 2 labeling was observed in the IPL (Fig. ). Inspection by confocal microscopy at higher magnification, however, showed that syntaxin 2 labeling was largely independent of syntaxin 1 labeling (Fig. ), although discrete syntaxin 2-positive puncta sometimes did colocalize with diffuse labeling for syntaxin 1. The differences in subcellular distribution between syntaxin 1 and syntaxin 2 suggest that these two isoforms are non-redundant, even though they are co-expressed by amacrine cells.
Figure 7 Syntaxin 2 shows little colocalization with syntaxin I. A–C: Syntaxin 2 labeling is present in numerous amacrine cells (A) and small puncta throughout the inner plexiform layer (IPL; arrows and arrowheads). Syntaxin 1 is widely expressed in amacrine (more ...)
The subcellular localization of syntaxin 2 suggested that it might be located in synaptic terminals lacking syntaxin 1, or that it might not be localized to a presynaptic compartment. To clarify this issue, double labeling for syntaxin 2 and known presynaptic markers was performed (Fig. ). There was little colocalization of labeling for syntaxin 2 and the conventional synapse marker synapsin I, although close apposition of labeled puncta was common (Fig. ). Likewise, double labeling for syntaxin 2 and the presynaptic active zone protein, bassoon, showed little colocalization but often showed apposition (Fig. ). Double labeling for syntaxin 2 and VGLUT1 confirmed that syntaxin 2 was not present in the ribbon synapses of bipolar cells, although puncta labeled for syntaxin 2 were often closely apposed to bipolar cell terminals (Fig. ). Double labeling for syntaxin 2 and synaptic vesicle protein 2 (SV2) using a pan-specific antibody that recognizes the SV2 isoforms at both ribbon and conventional synapses [23
], also showed little evidence of colocalization with syntaxin 2 (not shown). Together, these results indicate that syntaxin 2 was only rarely expressed in presynaptic conventional terminals and was not expressed presynaptically at bipolar cell ribbon synapses.
Figure 8 Syntaxin 2 is primarily localized outside presynaptic terminals. A–C: Syntaxin 2 is rarely localized to conventional presynaptic terminals in the inner plexiform layer (IPL). Syntaxin 2-positive puncta (arrows) do not colocalize with conventional (more ...)
Double labeling for syntaxin 2 and known ganglion cell and Müller glial cell markers showed no colocalization (microtubule-associated protein-1, MAP-1; glutamine synthetase, respectively. Not shown). These results, together with the results above, confirmed that syntaxin 2 labeling was restricted to amacrine cells and their processes in the IPL.
Syntaxin 4 is expressed in postsynaptic compartments in the OPL and IPL
Labeling for syntaxin 4 was the most restricted of all the isoforms examined (Figs. and ). Syntaxin 4 labeling was present in numerous small puncta in the OPL (Fig. ). These puncta were distributed individually in the distal OPL or in clusters more proximally (Fig. ), similar to the distribution of rod and cone terminals. In the IPL, a thin stratum of syntaxin 4-positive puncta was present at the INL/IPL border, with a few other syntaxin 4-positive puncta scattered more proximally in the inner portion of the IPL (Fig. ). Syntaxin 4 also was present in strings of small puncta in the INL that resembled the processes of interplexiform cell processes coursing vertically toward the OPL, however, the precise origins of these processes is unknown (Fig. ). Syntaxin 4 labeling also was found in the walls of blood vessels, and was especially prominent in relatively large diameter vessels along the inner surface of the retina (Fig. ).
Figure 9 Syntaxin 4 is highly restricted within both plexiform layers. A. Low magnification image of syntaxin 4 labeling. Labeling in the outer plexiform layer (OPL) is present in two sets of puncta. One set of small individual puncta in the distal OPL and a second (more ...)
Syntaxin 4 is present in horizontal cell processes post-synaptic to rod and cone terminals
To determine the relationship of the syntaxin 4 positive puncta to photoreceptor terminals in the OPL, we examined double labeling for syntaxin 4 and markers associated with photoreceptor terminals (Fig. ). Labeling for syntaxin 4 and post-synaptic density protein-95 KDa (PSD-95), which outlines the plasma membrane of rod and cone terminals [30
], showed that small syntaxin 4 positive puncta in the distal OPL were associated with rod terminals and the larger syntaxin 4-positive clusters in the proximal OPL were associated with cone terminals (Fig. ). Examination at high magnification revealed that syntaxin 4 positive puncta were enveloped by rod and cone terminals and often appeared as a doublet or a crescent (Fig. , respectively). Rod terminals enveloped one to four puncta; cone terminals enveloped numerous puncta. Patches of syntaxin-4 positive processes were often visible immediately below cone terminals and often connected to the puncta enveloped within the terminal (Fig. ), suggesting that syntaxin 4 labeling was associated with processes from second-order neurons. Labeling for syntaxin 4 and VGLUT1 were located adjacent to one another but did not colocalize (Fig. ), indicating that syntaxin 4 was not present within some highly localized pool of synaptic vesicles in the photoreceptor terminals. Syntaxin 4 labeling also did not colocalize with the synaptic ribbon markers kinesin 2 [31
] or bassoon [17
] (Fig. ; kinesin 2 shown), although the syntaxin 4-positive puncta were closely apposed to ribbons and often followed the contours of the ribbons. Together, these results indicate that syntaxin 4 is not present presynaptically in the rod or cone terminals, but is localized to processes of second-order neurons postsynaptic to the rod and cone terminals, including processes located within the ribbon synaptic complexes.
Figure 10 Syntaxin 4 labeling in the outer plexiform layer (OPL) is associated with processes post-synaptic to the photoreceptor terminals, not the terminals themselves. A–C: Double labeling for syntaxin 4 and the photoreceptor terminal marker PSD-95 show (more ...)
A series of double labeling experiments combining labeling for syntaxin 4 with known markers for specific subsets of second-order neurons or their synapses was performed to identify precisely which second-order neurons expressed syntaxin 4 (Fig. ). Double labeling for syntaxin 4 and calbindin, a marker for mouse horizontal cells [32
], showed extensive colocalization of syntaxin 4 and calbindin labeling in the horizontal cell processes, directly demonstrating that syntaxin 4 was present in horizontal cell processes at the ribbon complexes of rods and cones (Fig. ). Double labeling for syntaxin 4 and Go
α, a marker for ON-cone and rod bipolar cells and their dendrites [33
], showed no colocalization, indicating that syntaxin 4 was absent from the dendrites of rod and ON-cone bipolar cells (Fig. ). The dendrites of the ON-cone and rod bipolar cells often were seen terminating just below and between syntaxin 4 positive puncta in horizontal cell processes. This arrangement precisely matches the ultrastructural arrangement of ON bipolar cell dendrites and horizontal cell processes in the ribbon synaptic complexes. Double labeling for syntaxin 4 and peanut agglutinin (PNA), which labels the flat contacts on the base of cone terminals, showed that syntaxin 4 was not present in the dendrites of OFF-cone bipolar cells where they form flat contacts onto the cone terminals (Fig. ). Together, these results indicate that syntaxin 4 is specifically localized to the post-synaptic processes of horizontal cells occupying the lateral position at the ribbon synaptic complex, but not in the dendrites of ON bipolar cells in the ribbon complexes or OFF-cone bipolar cell dendrites at flat contacts.
Figure 11 Syntaxin 4 is localized to the tips of horizontal cell processes at ribbon synapses, but not to dendrites of ON- or OFF-bipolar cells. A–C: Syntaxin 4 is present in the tips of horizontal cell processes in the ribbon synaptic complexes of rods (more ...)
To better understand the potential functions of syntaxin 4 in the inner retina, double labeling for syntaxin 1, synapsin 1, SV2, and VGLUT1 was performed to characterize the relationship between the syntaxin 4-positive puncta and conventional and ribbon synaptic terminals in the IPL (Fig. ). Labeling for syntaxin 4 only rarely showed any colocalization with any of these presynaptic markers in either the INL or IPL, however syntaxin 4-positive puncta were often apposed to conventional and ribbon synaptic terminals. Consistent with these findings, syntaxin 4 labeling was often apposed to active zones labeled for bassoon, but did not colocalize bassoon itself (Fig. ). These results suggest syntaxin 4 in the IPL is located primarily in postsynaptic terminals.
Figure 12 Syntaxin 4 does not colocalize with syntaxin 1 or other presynaptic markers for conventional or ribbon synapses in the inner plexiform layer (IPL). A–C: Syntaxin 4-positive puncta (arrows) found in the inner nuclear layer (INL) and IPL do not (more ...)
Figure 13 Syntaxin 4 does not colocalize with the presynaptic active zone protein, bassoon, in the inner plexiform layer (IPL). A–C: Syntaxin 4-positive puncta in the IPL (arrowheads) do not show colocalization with presynaptic active zones containing bassoon (more ...)
Syntaxin 4 puncta in the IPL interact with dopaminergic and GABAergic amacrine cells
The plexus of syntaxin 4 puncta at the INL/IPL border is located precisely where dopaminergic amacrine cells and a population of GABAergic amacrine cells positive for CD15 have their plexes in the IPL [34
], suggesting that the syntaxin 4-positive puncta in the IPL might arise from or interact with these cell types. Double labeling for syntaxin 4 and tyrosine hydroxylase (TH), a marker for the dopaminergic amacrine cells, showed that the syntaxin 4 puncta co-stratified with processes from dopaminergic amacrine cells (Fig. ). Close examination of the arrangement of TH and syntaxin 4 labeling showed that the syntaxin 4 labeling was not present in processes arising from the dopaminergic cells, although there was contact between the processes of the dopaminergic amacrine cells and the syntaxin 4 puncta (Fig. ). In addition, syntaxin 4-positive puncta in the INL were often apposed to the cell body and short processes that emerged from the dopaminergic cells into the INL (Fig. ). A similar arrangement between the processes and cell bodies of the CD15-positive cells and the syntaxin 4-positive puncta in the distal IPL was observed (Fig. ). These results suggest that the dopaminergic and CD15-positive amacrine cells ramifying in the distal IPL may provide input to the syntaxin 4-positive puncta.
Figure 14 Syntaxin 4-positive puncta in the distal inner plexiform layer (IPL) do not arise from dopaminergic or GABAergic CD15-positive amacrine cells. A–C: Syntaxin 4-positive puncta form a plexus along the border of the inner nuclear layer (INL) and (more ...)
Double labeling for syntaxin 4 and the ganglion cell marker MAP-1 showed that syntaxin 4 labeling in the IPL was not associated with post-synaptic terminals of ganglion cell dendrites (Fig. ). Similarly, double labeling for syntaxin 4 and glutamine synthetase, a marker for Müller glial cells, showed no colocalization (Fig. ). Together with the results above, these findings suggest that the syntaxin 4-positive puncta in the INL and IPL do not arise from ganglion, bipolar or Müller cells, but are likely to arise from unidentified amacrine and/or interplexiform cells.
Figure 15 Syntaxin 4 in the inner plexiform layer (IPL) is not associated with ganglion cell dendrites or with Müller glial cell processes. A–C: Double labeling for syntaxin 4 and the ganglion cell dendrite marker, MAP 1. Syntaxin 4-positive puncta (more ...)