Reliable estimates of the quantum size in histaminergic neurons are not available. We have exploited two unusual opportunities in the fly's (Drosophila melanogaster) visual system to make such determinations for histaminergic photoreceptor synapses: 1) the possibility to microdissect successively from whole fly heads freeze-dried in acetone: the compound eyes; the first optic neuropils, or lamina; and the rest of the brain; and 2) the uniform sheaves of lamina synaptic terminals of photoreceptors R1–R6. We used this organization to count scrupulously the numbers of 30-nm synaptic vesicles from electron micrographs of R1–R6 profiles, and from microdissections we determined the regional contents of histamine in the compound eye, lamina, and central brain. Total head histamine averages 1.98 ng of which 9% was lost after freeze-drying in acetone and a further 28% after the brain was microdissected. Of the remainder, 71% was in the eye and lamina. Assuming that histamine loss from the tissue occurred mostly by diffusion evenly distributed among all regions, the overall lamina content of the head would be 0.1935 ng before dissection. From published values for the volumes of the brain's compartments, the computed regional concentrations of histamine are highest in the lamina (4.35 mM) because of the terminals of R1–R6. The concentration in the retina is ∼13% that in the lamina, suggesting that most histamine is vesicular. There are ∼43,500 ± 7,400 (SD) synaptic vesicles per terminal and, if all histamine is allocated equally and exclusively among these, the vesicle contents would be 858 ± 304 × 10−21 moles or ∼5,000 ± 1,800 (SD) molecules at an approximate concentration of 670 mM. These values are compared with the vesicle contents at synapses using acetylcholine and catecholamines.
The content of a synaptic vesicle, the type and amount of neurotransmitter, has traditionally been hard to determine. These difficulties do not diminish the value of making such determinations, however, either from biochemical studies on synaptic vesicle preparations or from studies on the size of the quantum, the synaptic vesicle's physiological counterpart ( Ceccarelli and Hurlbut 1980). The quantal content at cholinergic presynaptic terminals of the vertebrate neuromuscular junction has been determined, for example, by refined iontophoresis of acetylcholine that mimics the size of a single quantum ( Kuffler and Yoshikami 1975); but to determine the contents of a synaptic vesicle requires biochemical approaches on anatomically accessible populations of synaptic terminals, for which studies have progressed furthest on the electric organs of elasmobranchs ( Kelly et al. 1979). Attempts elsewhere in the nervous system are frustrated mostly by the mixture of different synapses in any one neuropil region.
Histamine is a neurotransmitter at the photoreceptors of many arthropod visual systems ( Callaway and Stuart 1989; Hardie 1987). Immunoreactivity to histamine, for example, is expressed in photoreceptors of the compound eyes of species as diverse as the cockroach Blaberus ( Pirvola et al. 1988) and horseshoe crab (Battelle et al. 1991). In the insect brain ( Nässel 1999), histamine was first confirmed as a neurotransmitter at the fly photoreceptor synapse, where iontophoresed histamine mimics light-evoked transmitter release ( Hardie 1987) and histamine receptor antagonists block light-evoked responses in postsynaptic neurons ( Hardie 1988). At such sites, histamine acts unusually as a fast transmitter, instead of at G-protein-coupled receptors as in vertebrates ( Roeder 2003; Stuart 1999), to transfer graded potentials (e.g., Juusola et al. 1995) by means of postsynaptic ligand-gated ion channels ( Gisselmann et al. 2001; Hardie 1989). Light-evoked release of histamine has also been reported from the compound eye of the fruit fly Drosophila melanogaster ( Sarthy 1991) but at rates that have yet to be quantified.
Well documented in the compound eye in flies ( Nässel et al. 1988; Pollack and Hofbauer 1991), histamine-like immunoreactivity is also expressed in three other locations. The first includes the other photoreceptor systems of the fly's head—the ocellus ( Nässel et al. 1988; Pollack and Hofbauer 1991) and a small extraocular system in flies ( Hofbauer and Buchner 1989; Yasuyama and Meinertzhagen 1999). The second is in wide-field neurons of the central brain ( Nässel et al. 1988; Pollack and Hofbauer 1991) that number ∼12 in each hemisphere of the Drosophila brain ( Pollack and Hofbauer 1991). The third site is at integumentary mechanoreceptors of the head epithelium of Drosophila ( Buchner et al. 1993). The overall histamine content of a Drosophila head is ∼2 ng ( Borycz et al. 2000). Of this, the compound eyes contribute the largest component. The total head content of histamine in eyeless sine oculis flies is only ∼30% that found in wild-type flies ( Borycz et al. 2002), suggesting that 70% of the head's histamine is associated with the compound eyes' photoreceptor neurons and their axon terminals in the two outer synaptic neuropils, the lamina and medulla.
The fly's visual neuropils are built on a modular principle ( Strausfeld and Nässel 1980). The first neuropil, or lamina, in particular comprises an array of cartridges. Each of these contains only ∼10 types of neuron, one each, and is innervated by the large cylindrical, vesicle-laden terminals of photoreceptor neurons R1–R6 ( Meinertzhagen and O'Neil 1991; Strausfeld and Campos-Ortega 1977). Photoreceptor synapses contain ∼60% of the total number of active zones between all cells; the remainder are found at presynaptic sites from lamina cells ( Meinertzhagen and Sorra 2001). Synapses of the photoreceptor terminals contribute most of the lamina's synaptic vesicles. These form a single population of clear profiles ( Meinertzhagen and O'Neil 1991) that is uniform, suggesting that each vesicle is a candidate quantum of photoreceptor histamine release; evidence for the existence of other photoreceptor neurotransmitters is lacking. Unlike R1–R6, which innervate the lamina, the axons of R7 and R8, the two central cells in each ommatidum, penetrate the lamina and pass by way of the external chiasma to form terminals in the more complex distal medulla ( Fischbach and Dittrich 1989).
Given the lamina's anatomical features, we have sought a simple method to isolate this neuropil from other regions of the fly's brain and to make repeatable determinations of its histamine content. Such determinations are necessary before physiologically more important parameters, such as data on histamine release, can be evaluated. We have developed a sensitive method using electrochemical detection in association with high-performance liquid chromatography (HPLC) to detect histamine in Drosophila heads ( Borycz et al. 2000, 2002). This method provides a means to determine the histamine content of the fly's head, but only for the entire head and not for individual compartments within its brain. For the latter purpose, we have modified a microdissection method applied to the brain after freeze-drying heads in acetone ( Fujita et al. 1987).
Wild-type Oregon R Drosophila melanogaster maintained in laboratory cultures at 23°C in a 12-h light:dark regime on standard cornmeal medium were used for all histamine determinations.
To dissect the different components of the visual system and brain of Drosophila, we used a previously reported method of freeze-drying the head from acetone ( Fujita et al. 1987). Fly heads were fixed on ice for 5 h in 4% 1-ethyl-3-(3-dimethylamino-propyl) carbodiimide (E-7750: Sigma, St. Louis, MO) and immediately freeze-dried in acetone, which was previously chilled to -80°C. After 10 days in a freezer at −80°C, the samples were placed on filter paper to let the remaining acetone evaporate and were then stored in a desiccator over Drierite dessicant and evacuated by a tap aspirator. Different components were dissected from these brains using mounted tungsten needles, according to the method of Fujita et al. (1987) as illustrated in Fig. 1.
Histamine determinations were performed using HPLC with electrochemical detection as reported for Drosophila ( Borycz et al. 2000, 2002). This method uses o-phtaldialdehyde-mercaptoethanol to convert (derivatize) histamine to a form having an electrochemical potential that is measurable by electrochemical detection. The method reflects a histamine recovery rate >90% and was previously used to determine average head histamine contents from samples of 50 heads ( Borycz et al. 2002). Determinations from samples of as few as 20 heads became possible in the present study using improved methods, however, so that each sample was made from a derivatized homogenate of between 20 and 50 fly heads. Details of sample numbers are given in results.
For microdissection to be useful in determining the regional content of brain histamine, we first needed to ascertain how much histamine was retained in the head after freeze drying. The retention levels when whole heads were simply frozen in acetone were very low, <20%, presumably because histamine dissolved in the acetone and diffused away from the head. We therefore first fixed the head in carbodiimide, a fixative that has previously been used to conjugate histamine to succinylated keyhole limpet hemocyanin to raise a widely used polyclonal antiserum against histamine ( Panula et al. 1988). The latter detects histamine in frozen sections of insect brains fixed in carbodiimide ( Nässel et al. 1988; Pirvola et al. 1988; Pollack and Hofbauer 1991), suggesting that carbodiimide in some way fixes histamine in brain neuropils. Retention rates from microdissected, carbodiimide-fixed brains were sufficient to allow successful determinations.
To reveal the distribution of histamine in the Drosophila brain, frontal sections of the fly's head were cut and immunolabeled using reported methods ( Buchner at al. 1993; Pollack and Hofbauer 1991). After fixation on ice with 4% 1-ethyl-3-(3-dimethylamino-propyl) carbodiimide, heads were cryoprotected in 25% sucrose and cut on a cryostat (Reichert-Jung 2800 Frigocut) at 10 μm thickness. The sections were incubated with a rabbit polyclonal antibody (PAN19C; Immunostar, Stillwater, MN) at 1:500 and a Cy3-conjugated goat anti-rabbit secondary antibody (Jackson ImmunoResearch, Bar Harbor, ME) at 1:400. Images of immunolabeled sections were collected by confocal microscopy (Zeiss LSM410) using a ×40/1.30 immersion objective.
Synaptic vesicle numbers
To calculate the number of synaptic vesicles per R1–R6 terminal in the lamina, cross-sections of lamina cartridges were captured digitally from material prepared for electron microscopy (EM) using methods reported previously ( Meinertzhagen 1996). Counts of the ∼30-nm-diameter profiles of synaptic vesicles were made from ultrathin sections 50- or 65-nm thick. To correct for the caps of vesicle profiles, but given that the shapes of synaptic vesicles approximate simple spheres, Abercrombie's correction was applied where n = raw vesicle profile count; Nc = corrected vesicle number; t = section thickness (50 or 65 nm); and d = mean vesicle diameter (30 nm).
All values, for both histamine content and vesicle numbers, were compiled as means ± SD of the mean values for a number of samples, ≥12 for histamine determinations, and 5 for synaptic vesicle counts, as reported in greater detail in RESULTS. To compare synaptic vesicle counts, we first made an unweighted means ANOVA, followed by a Tukey HSD test, using software (Systat 5.2.1).
Histamine contents of fly heads
The value for the head content of histamine is not entirely fixed but increases if Drosophila are first given a 4% aqueous solution of glucose to drink (J. Borycz, unpublished observations). It can also be increased by feeding flies with a solution containing either 0.5% histamine or 5% l-histidine, the precursor for histamine synthesis ( Burg et al. 1993), or 0.5% of the major fast metabolite of histamine, β-alanyl histamine, or carcinine ( Borycz et al. 2002). In the experiments reported in the current study, determinations were therefore made on samples from flies with a carefully monitored dietary history, either 50 fly heads or, using improved methods, from 20 fly heads.
Histamine contents of the components of the Drosophila brain
Even though we were able to microdissect the compound eyes and laminas successively from the central brain, using the microdissection of freeze-dried heads, we could not reliably separate the second optic neuropil, or medulla, from the central brain. The medulla receives innervation from the terminals of R7 and R8 (Fig. 2). The histamine contents of the terminals of photoreceptors R7 and R8 were therefore lumped with those for the ∼12 central brain neurons.
Seventeen histamine determinations of carbodiimide-fixed freeze-dried heads, 14 from 50-head samples and 3 from 20-head samples, were made (Table 1). The amount of histamine retained after freeze drying whole heads that were previously fixed in carbodiimide averaged 1.802 ng, ∼91% of the total histamine in fresh heads, which is 1.98 ± 0.15 ng ( Borycz et al. 2000).
We microdissected 12 batches each of 20 carbodiimide-fixed freeze-dried fly heads, to isolate the brain's three component (Table 1). Incomplete or imperfect dissections (initially 50% but improving to 15%) were discarded. After we microdissected the head to isolate the brain's three component regions, an average of 70% of the freeze-dried head total histamine remained. The 30% unaccounted for could have reflected the presence of histamine in nonneural structures of the head (cuticle, fat body, muscle, etc), although these have not hitherto been reported; or, more likely, two main sources of loss: the loss of neural tissue during microdissection; or the loss of histamine from neural tissue into the acetone. Because each compartment was dissected consecutively from the same head, the totals were cumulative, and lost head tissue could be only the traces of tissue remaining either on the glass dissecting substrate or on the dissection needles themselves. We estimate that these quantities were small compared with 30%. Of the brain histamine surviving microdissection, an average cumulative total of 1.254 ng, 52% occurred in the retina and a further 20% in the lamina, for a total of 71%, roughly the proportion of histamine reduced in eyeless sine oculis flies relative to wild type ( Borycz et al. 2000). All histamine contents are given per head, that is, for two eyes, two laminas, and the whole remainder of the brain including the distal medulla of each side, which is innervated by the histamine-immunoreactive terminals of R7 and R8 ( Pollack and Hofbauer 1991).
Regional concentrations of histamine
Based on the histamine content of the brain's three compartments that we could successfully microdissect, the compound eyes, the lamina, and the rest of the brain, we can calculate the concentrations of histamine from published values for the respective volumes of these regions (Table 1). Volumetric determinations are reported from light microscopic determinations by Hauser (1975), using reduced silver preparations, and by Barth et al. (1997), using autofluorescence profiles, also on paraffin wax sections.
To calculate the concentration of histamine, we first needed to estimate that portion of the histamine lost from each of the separate compartments. If we assume that the loss, 30%, was distributed equally from each compartment, we conclude that the overall lamina content of the head was 0.387 ng before dissection. The volume of the lamina neuropil beneath the seeing eye of a fly reared under light/dark conditions is reported to be ∼4 × 105 μm3 ( Barth et al. 1997), giving a concentration of 0.48 μg/μl or 4.35 mM. The volume of the retina is reported to be ∼0.008 mm3, giving an overall retinal concentration of 0.574 mM. This is attributable to the photoreceptor somata, and thus ∼13% that in the lamina, where the terminals of R1–R6 reside.
Both of these concentrations are for the entire compartment that includes not only photoreceptors but also other cells, some of which might also contain histamine. For the ommatidia in the retina, the somata of R1–R6 are surrounded by pigment cells, and the separate volumes are apparently not reported. For the lamina the terminals of R1–R6 occupy as much as 73% of the combined volume of the cartridge and glia ( Barth et al. 1997) (Fig. 3). Thus if the lamina's histamine concentration of 4.35 mM was concentrated entirely within the terminals of R1–R6, their overall histamine concentration could be nearly 6 mM. Some histamine may be taken up by the lamina's glia, however, at two cell types. First, epithelial glia (Fig. 3) are the site of histamine metabolism regulated by the product of the gene ebony ( Borycz et al. 2002), for which histamine uptake is required; and, second, a band of immunolabeling in the proximal lamina is reported ( Borycz et al. 2002) in a position, illustrated in Fig. 2, possibly corresponding to the marginal glia ( Eule et al. 1995).
Number of synaptic vesicles in a R1–R6 photoreceptor terminal
To correlate the regional content of histamine with a population of identified synaptic terminals, we chose the lamina (Fig. 4A). This choice was based on four reasons: the largest cell type in the neuropil is the R1–R6 terminal ( Barth et al. 1997); the number of terminals is defined; these terminals alone are reported to contain histamine ( Pollack and Hofbauer 1991); and, finally, we could microdissect the lamina to obtain the histamine content of all terminals. We next undertook counts of the synaptic vesicles in the terminals of R1–R6, which are summarized in Table 2. These formed clear, round or pleiomorphic profiles ∼30 nm in diameter, which coexisted together with other synaptic organelles (Fig. 4B). It is perhaps surprising that counts of vesicle profiles are available in the literature only from a single unpublished study ( Hauser 1975).
Counts of synaptic vesicle profiles were made on the cross-sectioned profiles of R1–R6 terminals, in tangential sections of the laminas of five wild-type flies. Four of these were from Oregon-R cultures, flies 1–3 from one stock and a fourth fly 4 from a second. In these four, we counted from R1–R6 profiles that contained the middle regions of the lamina's depth from cartridges sampled across the entire area of sections that did not contain the chiasma. Thus our sample contained cartridges that were sectioned more proximally, in the center of the section, and cartridges sectioned more distally, at its edges. The sampled region was ∼10 cartridges wide, the maximum number that were transversely sectioned, and thus contained ∼80 of the lamina's >750 cartridges. A fifth fly was taken from the wild, and counts of its synaptic vesicles were made from the six R1–R6 terminals from a single, serially sectioned cartridge from which the synaptic organization has previously been reported ( Meinertzhagen and O'Neil 1991; Meinertzhagen and Sorra 2001). Raw counts of profiles were corrected by the Abercrombie formula (1946), and the density of vesicles per unit area of R1–R6 cross-section profile area calculated.
The five corrected mean vesicle counts and vesicle density counts exhibited relatively wide SDs, ∼12.5% of the mean (Table 2). The corrected counts for fly 2, in particular, were lower, and differed significantly from both fly 4, the sole Oregon R fly, and fly 5 (both P < 0.05), but not from the two others in the same stock. This suggests that some individual physiological condition, rather than genetic background, gave rise to the variation in vesicle numbers.
For the fifth fly, we used the depth of the lamina previously calculated from the length of the synaptic terminals of R1–R6 ( Meinertzhagen and O'Neil 1991) to calculate the total number of synaptic vesicles per terminal. This averaged 38,432 vesicles per terminal. Assuming a similar lamina depth for the other four flies, we obtained corresponding values ranging from 33,326 to 51,134 synaptic vesicles per terminal. The lowest value was for fly 2, as in the preceding text, and was again significantly different (P = 0.015).
The average corrected number of vesicles calculated for all five flies was 43,550 with a SD of ±7,434, 17.1% of the mean (Table 2). The corresponding mean cross-sectional area of the terminals for the four Oregon-R flies was 2.2036 μm2, giving a mean R1–R6 terminal volume of 63.90 μm3. The packing densities of synaptic vesicles, per μm2 of section and per μm3 of terminal volume, showed much greater variation than the raw vesicle numbers, with respective SDs 39.6 and 31.3% of the corresponding mean values. This difference implies that the vesicle population is more closely regulated than the girth of the terminal containing it rather than that co-variation in both variables, synaptic vesicle count and R1–R6 terminal profile girth, produced a constant vesicle packing density.
Each vesicle has a diameter of 30 nm and thus a volume of 1.414 × 10−5 μm3. Thus the mean vesicle numbers represent about 0.96% of the cytoplasmic volume of the terminal. These numbers compare with an uncorrected value of 93,000 vesicles from profile counts calculated by Hauser (1975) or ∼5.2% of the terminal's volume calculated from data recorded by that author. The difference between these percentage values lies in the uncorrected larger number of synaptic vesicles and the smaller terminal volume recorded by Hauser (1975).
Histamine content of a synaptic vesicle in a R1–R6 photoreceptor terminal
Finally we used the lamina's histamine content and the synaptic vesicle counts presented in the preceding text to derive an approximate estimate of the amount of histamine per vesicle (Table 3). Given that most neurotransmitter is segregated in the synaptic vesicles and provided that all synaptic vesicles are equally charged with neurotransmitter, this estimate would then give an approximate idea of the histamine content of each vesicle.
The total number of synaptic vesicles in all lamina terminals of R1–R6 depends on the number of cartridges, each with a normal complement of six terminals, the overall average for all cartridges. The number of cartridges is most readily estimated from the number of ommatidia that innervate them ( Braitenberg 1967); this has been estimated to average 776 for a sample of wild-type eyes ( Ready et al. 1976). The lamina's complement of vesicles is the simple product of all these values or 2.028 × 108. The mean histamine content for a single lamina was 0.193 ng or 1.740 × 10−3 nmole and thus contained ∼1.05 × 1012 molecules. Distributed equally among all vesicles, each would contain 8.58 ± 3.04 × 10−21 moles: 8.58 × 10−6 fmole or 5,178 ± 1,830 (SD) molecules.
Here, we report the histamine contents, their compartments and corresponding concentrations for the brain of D. melanogaster, together with synaptic vesicle counts for the terminals of R1–R6 in the lamina, and a final estimate of the amount of histamine per vesicle, ∼5,000 molecules. Each of these quantities includes variations and other uncertainties, which we will acknowledge in turn in evaluating our data. Those for the content and concentration of histamine are illustrated schematically in Fig. 5.
Our determination of vesicle histamine content is totally straightforward, differing from most approaches in vertebrate systems chiefly in not depending on the technical problems associated with isolating synaptic vesicle. The opportunity to derive an estimate of the quantum content of the fly's photoreceptor synaptic terminal rests on two particular features of the fly's lamina. The first is the favorable geometry of the lamina and the opportunity to dissect it cleanly from both the overlying retina and the underlying medulla. Although technically demanding, the efficient microdissection of the optic neuropils into batches of sufficient number to subject to histamine determinations has yielded determinations with acceptably low levels of variation. The second feature is the near perfect geometrical regularity of the terminals of R1–R6. This regularity and the arrangement of terminals in sheaves enable accurate counts of synaptic organelles ( Meinertzhagen 1996). Even so, our counts of synaptic vesicles exhibit some variation, having a SD that is 17% of the mean. Possibly unintended variation in section thickness could have contributed some differences in the raw counts of vesicle profiles.
Errors in determinations
The determinations of histamine in the entire freeze-dried head exhibit variations that are small, with a SD only ∼2.45% of the mean (Table 1). After microdissection, however, the errors for different brain regions are larger, 26% for the eye, 31% for the lamina, and 28% for the rest of the brain. The cumulative error is smaller again, however, the SD of all determinations being only 16% of the mean. Thus losses in one compartment appear to offset gains in the adjacent compartment, implicating microdissection as a major source of variation between sample determinations. We were unable to evaluate microscopically how cleanly each compartment broke from the surrounding one, however, to confirm the errors in dissection, because the tissue was freeze-dried and its microscopic structure lost.
The completeness of the compartmental determinations of histamine depends critically on retaining ∼90% of whole-head histamine after carbodiimide fixation. This retention rate highlights an apparent paradox, however: carbodiimide fixation stabilizes histamine within the different compartments during microdisseection yet does not prevent the liberation of free histamine during derivatization. Carbodiimide acts by forming a peptide bond between a carboxyl and amino group and thus does not directly bind histamine ( Panula et al. 1988). Presumably therefore it helps retain cellular histamine by forming peptide bonds among surrounding proteins, thereby impeding histamine diffusion. Regardless, histamine must be liberated during derivatization to yield detectable amounts that are similar to those determined in fresh unfixed heads.
The regional concentrations of histamine incorporate additional errors attributable to the computed volumes of the different brain regions. Accurate data are lacking for most volumes, values for which derive from paraffin wax section series ( Barth et al. 1997; Hauser 1975) and thus incorporate errors through tissue shrinkage. The volumes used in Table 1 are not corrected for this shrinkage, estimated to be ≥15% in linear extent ( Strausfeld 1976) and thus >40% in volume. A recent wholemount study that does not rely on histological embedment gives accurate volumetric data based on large sample numbers but unfortunately does not include the lamina ( Rein et al. 2002). Comparisons between these different studies are therefore problematic, but, as an example, the cumulative neuropil volume of the central brain calculated from wholemount data ( Rein et al. 2002) is 0.00427 mm3, whereas from wax sections the volume of the entire brain minus the volume of the lamina, which thus includes cortices as well as neuropil regions, is 0.00636 mm3 ( Hauser 1975). For the lamina, the volumes are 10 times less, either ∼0.4 × 10−3 mm3 ( Barth et al. 1997) for the neuropil proper or 0.64 × 10−3 mm3 including the cortex ( Hauser 1975). These errors do not enter into our calculations of synaptic vesicle content.
Our estimates of the amount of histamine per synaptic vesicle in R1–R6 incorporate the errors in histamine determinations and also convolve errors in synaptic vesicle numbers, which have a standard error of ∼17% of the mean. Two sources of that variation are sampling errors and the physiological condition of the cells. In the first case, we made raw vesicle profile counts from several flies but only from a single section from each, and section thickness is difficult to control. In the second case, diffusion fixation for EM may not capture the complete vesicle population without change. Unpublished evidence suggests, for example, that the R1–R6 vesicle pool stabilizes only 6 or 7 min after the onset of aldehyde fixation at 4°C (MacIntosh 1996), whereas vesicle endocytosis takes no more than a minute ( Betz and Wu 1995). Thus the vesicle cycle should still be active in the fixative.
In addition to these computational errors, it is not clear how many of the vesicles, which are distributed in a large population throughout the cytoplasm of the terminals, are pumped with histamine. At glutamatergic synapses vesicles are pumped early in their cycle ( Prior and Clague 1997), but evidence is lacking for R1–R6. It has been proposed that the capitate projections (Fig. 4B) are sites of neurotransmitter recycling ( Fabian-Fine et al. 2003), so that possibly some vesicles close to the heads of the capitate projections are empty. Moreover, not all histamine need be vesicular. Cytoplasmic neurotransmitter concentration is however low compared with that in synaptic vesicles, perhaps in the ratio 1:100 for cholinergic synapses ( Parsons et al. 1993), supporting which the histamine concentration in the photoreceptor's soma in the retina is only 13% of that in its lamina terminal.
Quantum content of a histamine vesicle
Neglecting the nonvesicular distribution of histamine, and assuming that most histamine is sequestered in synaptic vesicles rather than the cytosol, we can calculate the histamine content per vesicle in a terminal of R1–R6. The calculations we present in Table 3 suggest that each vesicle contains 8.58 × 10−6 fmole or 5,178 molecules within a spherical envelope having an inside diameter no greater than 28–30 nm. Thus the vesicle contains 0.858 × 10−5 fmole of histamine in a volume of 1.277 × 10−5 fl, so that the histamine concentration in the interior must be of the order of 670 mM. This exceeds considerably the osmolarity of hemolymph in which synapses function, which for larval Drosophila ranges between 360 and 396 mosM ( Stewart et al. 1994) but is only half the osmolarity of the modified Karnovsky primary fixative used to preserve photoreceptor synaptic vesicles ( Meinertzhagen 1996), which exceeds 1,400 mosM ( Rasmussen 1974). The superior EM preservation provided by hypertonic fixatives is well established.
These estimates are approximate. The combined SDs of the two main determinations for this estimate, histamine content and vesicle number, indicate that variation is ∼36%, but there are other uncertainties as given in the preceding text. There is no evidence that photoreceptor synaptic vesicles contain any other neurotransmitter, so that all vesicles should contain, or be destined to contain, histamine. The effects of unfilled vesicles would be to underestimate the histamine content, whereas the effects of appreciable amounts of cytosolic histamine would be to overestimate the vesicle content. Our evidence from the retinal concentration suggests that this overestimate could be as much as 13%, giving a vesicle content more like 4,600 molecules.
It is interesting that the contents compare to similar values for synaptic vesicles at other synapses. Preparations from the electric organ in the elasmobranch Narcine contain ∼47,000 molecules of acetylcholine within each 80-nm-diam vesicle ( Wagner et al. 1978). The quantal content of acetylcholine at the frog and snake neuromuscular junction is said to comprise <10,000 molecules ( Kuffler and Yoshikami 1975). Synaptic vesicles at the frog neuromuscular junction average only 42 nm in diameter ( Heuser and Reese 1981) and are thus scarcely one-seventh the volume of the larger cholinergic vesicles of the electric organ. On the other hand, vesicles at the frog neuromuscular junction have a volume that is three times larger than the 29-nm vesicle interiors in R1–R6 terminals.
For the vesicles at noncholinergic synapses, those at rod terminals in the vertebrate retina are estimated to contain 2,000 molecules of glutamate in a vesicle 28 nm in diameter ( Rao-Mirotznik et al. 1998). These values are not dissimilar to those for Drosophila R1–R6 terminals. In our data, we find a high intravesicular concentration of histamine, ∼0.67 M, comparable with the reliable values reported for chromaffin cells. For the latter, patch-amperometry data yield catecholamine concentrations of ≥0.7 M, regardless of granule size ( Albillos et al. 1997), similar to the value of 0.6 M previously obtained from biochemical studies ( Phillips 1982). These concentrations, and ours, are both higher than for the intravesicular concentration in rod vesicles on which Rao-Mirotznik et al. (1998) base their estimates of quantum size in rod vesicles, however. Thus reported values for intravesicular glutamate concentration fall between 60 ( Burger et al. 1989) and 210 mM ( Riveros et al. 1986), much lower than we see, but derive from studies on vesicle preparations, whereas our value is from intact tissue. There are apparently no prior reliable estimates of vesicle contents at histaminergic synapses for which our study is thus a precedent.
If we accept the vesicle's mean histamine contents, we can go on to estimate the total histamine released from an R1–R6 terminal from published estimates of vesicle release rates. Fly photoreceptors have high-gain synapses with high rates of tonic neurotransmitter release (e.g., Uusitalo et al. 1995). For example, large fly species have been estimated to release ∼100 vesicles per release site per second ( Laughlin and de Ruyter van Steveninck 1996), equivalent (if quantal size is the same as in Drosophila) to ∼9 × 10−4 fmole or 0.5 to 15 × 106 molecules/s. Were all 50 release sites at a Drosophila R1–R6 terminal ( Meinertzhagen and Sorra 2001) to release at this rate, without compensatory endocytosis, the terminal's 43,500 or so total vesicle population would deplete at ∼5,000 vesicles/s. Thus either vesicle recycling either by “kiss-and-stay” ( Südhof 2004) or “kiss-and-run” ( Ceccarelli et al. 1973; Valtorta et al. 2001) at the release site or by clathrin-mediated endocytosis at capitate projections ( Fabian-Fine et al. 2003), is sufficient to compensate, or release rates in Drosophila are in fact usually less.
This work was supported by National Eye Institute Grant EY-03592 and by Canadian Institutes of Health Research Grant ROP-67480 to I. A. Meinertzhagen.
We thank J. A. Horne for help with quantitative aspects of this study and with the preparation of Fig. 1. Dr. X. Sun immunolabeled and collected the image stack of the wholemounted brain shown in this figure.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2005 by the American Physiological Society