Proc. Nat. Acad. Sci. USA Vol. 72, No. 9, pp. 3575-3579, September 1975 Cell Biology Brain tryptophan hydroxylase: Purification of, production of antibodies to, and cellular and ultrastructural localization in serotonergic neurons of rat midbrain (immunocytochemistry/microtubules/raphe nucleus/serotonin/tryptophan 5-monooxygenase) TONG HYUB JOH, TADAHIRO SHIKIMI, VIRGINIA M. PICKEL, AND DONALD J. REIS Laboratory of Neurobiology, Department of Neurology, Cornell University Medical College, 1300 York Avenue, New York, N.Y. 10021 Communicated by Seymour S. Kety, June 19, 1975

ABSTRACT Tryptophan hydroxylase [EC 1.14.16.4; L- droxylase (11-15), a closely related enzyme which subserves tryptophan, tetrahydropteridine:oxygen oxidoreductase (5- the biosynthesis of catecholamine neurotransmitters, has en- hydroxylating)], the enzyme catalyzing the rate-limiting step in the biosynthesis of serotonin, was purified 79-fold from couraged us in this study to attempt purification and local- the region of the raphe nucleus of rat midbrain by sequential ization of tryptophan hydroxylase in brain. column chromatography and disc-gel electrophoresis. In elec- trophoresis three bands were distinguished, A, B, and C, which, when separated and submitted individually to elec- MATERIALS AND METHODS trophoresis, reproduced the same three bands. Bands A and C were enzymatically active and inhibited by para-chloro- phenylalanine. Antibodies produced to each of the three Materials. L-Tryptophan, dithiothreitol, catalase, and bands crossreacted by immuno double diffusion and electro- pyridoxal-5'-phosphate were purchased from Sigma Chemi- phoresis with each other and homogenates of raphe nuclei; cal Co.; 2-amino-4-hydroxy-6,7-dimethyl-5,6,7,8-tetrahy- they completely inhibited enzyme activity only of trypto- dropteridin from Aldrich Chemical Co.; NCS-Tissue Solubil- phan hydroxylase. Tryptophan hydroxylase was localized by izer from Amersham-Searle Corp.; L-[1-'4C]tryptophan light and electron immunohistochemistry to serotonin neu- from New England Nuclear Corp. Sepharose 4B and Sepha- rons of the raphe. Ultrastructurally, in cell bodies, the en- zyme was distributed in cytoplasm and in association with dex G-200 were obtained from Pharmacia Fine Chemical, endoplasmic reticulum and Golgi apparatus. In dendrites Inc., and acrylamide H.P. and N,N'-methylene bisacrylam- and axons, it was associated with microtubules. Tryptophan ide H.P. from Canalco, Inc. hydroxylase in brain is only neuronal and cytoplasmic, exists Tryptophan hydroxylase was purified from the region of in multiple forms, and is associated with microtubules, the raphe nuclei of the midbrain of 260 male Sprague- suggesting it may be transported from sites of synthesis in Dawley rats dissected according to the procedure of Degu- cell body into axons. chi et al. (16). Aromatic L-amino-acid decarboxylase (EC The enzyme, tryptophan hydroxylase [EC 1.14.16.4; L-tryp- 4.1.1.28; aromatic-L-amino-acid carboxy-lyase) was partially tophan, tetrahydropterine:oxygen oxidoreductase (5-hydrox- purified from hog kidney of DEAE-Sephadex column chro- ylating)] catalyzes the initial and probably rate-limiting step matography by the method of Christenson et al. (17). in the biosynthesis of the neurotransmitter serotonin (5HT) Enzyme Assay. Tryptophan hydroxylase activity was as- (1, 2). In brain, the enzyme has been presumed to be con- sayed by a modification of the method of Ichiyama et al. tained only within those specific neurons which synthesize, (10), in which the rate of conversion of L-[1-'4C]tryptophan store, and release 5HT (3). The evidence is indirect and to '4CO2 was measured. The assay mixture contained 200 Al based on the reasonable assumption that those neurons in of enzyme preparation, 50 ,l of 1 M Tris-acetate buffer, pH which 5HT can be visualized by their specific histofluores- 7.6, 10 Al of 30 mM dimethyltetrahydropteridin in 120 mM cence (4, 5) should contain the enzyme. Consistent with the dithiothreitol, 66 units of catalase in 10 Al, 10,l of 1 mM surmise are biochemical data demonstrating the presence of ferrous ammonium sulfate, and 20 Al of 150 ,M L-[1- high levels of tryptophan hydroxylase activity in brain re- 14C]tryptophan. The mixture was incubated for 30 min at gions containing the cell bodies of 5HT neurons (3) or richly 370, and 20 ,l (30 nmol/min) of aromatic L-amino-acid de- innervated by their processes (3). It has also been proposed, carboxylase and 40 Al of 2.5 mM pyridoxal-5'-phosphate on indirect evidence, that tryptophan hydroxylase exists in were added. The assay tube was then covered by a rubber different forms, differing either in susceptibility to inhibi- vial cap (Kontes Glass Co.) in which was suspended a plastic tion by drugs (6-8) or possibly, in their relationship to sub- well containing a folded filter paper wick and 250 ,l of NCS cellular organelles (3). Tissue Solubilizer. After an additional 30 min of incubation Direct information of the biochemical structure and re- at 37', 100 Ml of 30% perchloric acid was injected through gional and subcellular localization of the enzyme has been the rubber cap and 14CO2 was trapped at 37° for 90 min. restricted by limited success (9, 10) in obtaining purified The plastic well was removed and the radioactivity was de- tryptophan hydroxylase suitable for the development of spe- termined in 10 ml of Bray's solution by liquid scintillation cific antibodies for immunohistochemistry. Recently the de- spectrometry. Activity was expressed as nmol of CO2 velopment in our laboratory of techniques for the purifica- formed per 60 min. One unit of tryptophan hydroxylase ac- tion and immunohistochemical localization of tyrosine hy- tivity is defined as that amount of enzyme which produces 1 nmol of CO2 per 60 min. Protein was determined by the method of Lowry et Abbreviation: 5HT, serotonin. al. (18), except in monitoring column Requests for reprints and all correspondence to: Dr. Tong H. Joh, effluents, when the absorbance at 280 nm was taken as a Department of Neurology, Laboratory of Neurobiology, Cornell measure of protein concentration. University Medical College, 1300 York Avenue, New York, N.Y. Electrophoresis. Polyacrylamide disc gel electrophoresis 10021. was carried out by the method of Davis (19) and Ornstein 3575 Downloaded by guest on October 3, 2021 3576 Cell Biology: joh et al. Proc. Nat. Acad. Sci. USA 72 (1975)

(20) using 7% gel, 0.4 M glycine-Tris buffer, pH 8.2 contain- mM dithiothreitol using a Polytron tissue homogenizer. ing 2 mM dithiothreitol, and 3 mA per tube. Amido black The same buffer was used for Sepharose and Sephadex was used to stain the protein on the gel. To assay enzyme ac- column chromatography. The homogenate was centrifuged tivity of different bands, gels were sliced into pieces of 2 at 39,000 X g for 30 min, and powdered ammonium sulfate mm length and placed in 250 ,l of 0.3 M Tris-acetate buff- was added to the supernatant to 80% saturation. The enzyme er, pH 7.0 containing 2 mM dithiothreitol to elute enzyme. was further purified by three sequential chromatographic The tubes were allowed to stand overnight at 40, and the en- steps: In the first step, the enzyme was passed through a zyme activity was determined in the buffer after the gel Sepharose 4B column achieving 2.8-fold purification of en- slices were removed. zyme and 80% recovery of activity. This step removed, in Production of Antibodies. Purified tryptophan hydroxyl- the void volume, an enzymatically inactive peak containing ase was subjected to polyacrylamide disc gel electrophoresis. substances of high molecular weight and having a high spec- Segments containing active enzyme were pooled from 24 trophotometric reading at 260 nm, indicating the presence gels and homogenized in a glass tube with 2.0 ml of 0.9% so- of compounds with a relatively high content of nucleic dium chloride. An equal volume of complete Freund's adju- acids. If this fraction was not removed, subsequent purifica- vant was added; the mixture was thoroughly emulsified and tion procedures were ineffective because the enzyme would injected subcutaneously into white New Zealand rabbits. aggregate into a high-molecular-weight form which was The injections were repeated every 3 weeks for a period of 9 then inseparable from other proteins by chromatography or weeks. The rabbits were bled 5 days after the final injection. electrophoresis. Immunoglobulin (IgG) was precipitated from the serum at Active fractions were then passed through a hydroxyapa- 50% saturation with ammonium sulfate. Immunodiffusion tite column by stepwise elution of protein using 0.0625, and immunoelectrophoresis were carried out by the methods 0.125, and 0.25 M potassium phosphate buffer, pH 6.8, con- of Ouchterlony (21) and Scheidegger (22), respectively. taining 2 mM dithiothreitol. The highest specific activity Immunohistochemical Localization of Tryptophan Hy- was obtained by elution with 0.125 M buffer. The final step droxylase By Light and Electron Microscopy. Tryptophan consisted of Sephadex G-200 column chromatography in hydroxylase was localized immunohistochemically in rat which a 2.4-fold purification was obtained. The specific ac- brain by use of the peroxidase-antiperoxidase method of tivity of the purified enzyme was 45 units/mg of protein, Sternberger et al. (23). The techniques were identical to which represented a 79.2-fold purification of the enzyme those used by us for localization of tyrosine hydroxylase, and from the 39,000 X g supernatant. the methods are extensively detailed elsewhere (13, 15). After each chromatographic step the enzyme was subject- In summary, rats were anesthetized with pentobarbital ed to polyacrylamide gel electrophoresis. The number of (50 mg/kg) and perfused through the aorta with either 4% protein bands detected on the gels after each step was not paraformaldehyde in phosphate buffer, pH 7.3, or with di- substantially reduced despite an increase in specific activity. lute Karnovsky fixative (1% glutaraldehyde and 1% parafor- Electrophoresis of the enzyme after the final chromato- maldehyde in 0.1 M cacodylate buffer at pH 7.15). For light graphic step, the G200 effluent (Fig. 1), produced three microscopy, the formalin-perfused brain was post-fixed for heavily stained (major) bands designated A, B, and C (in re- 6 hr with picric acid/formalin, washed overnight in phos- verse order of electrophoretic mobility) and several lightly phate buffer, embedded in polyethylene glycol, and sec- stained (minor) bands. Only two of the major bands, A and tioned on a rotary microtome at 5 ,m. For electron micros- C (Fig. 1), exhibited enzyme activity, band C having the copy, perfusion of the brain with Karnovsky's solution was greater specific activity. The remaining band, B, was inac- followed by fixation for 2 hr in 2% glutaraldehyde parafor- tive. Enzyme activity of both active bands, A and C, were maldehyde. Sections of 20 ,um were taken through the mid- completely inhibited by p-chlorophenylalanine, an inhibitor brain on a vibrating microtome (Vibratome) with the tissue of tryptophan hydroxylase (6-8). immersed in 0.1 M cacodylate buffer. Sections prepared for Each major band, A, B, and C, was then isolated from 33 light or electron microscopy were sequentially incubated gels, eluted, and subjected, by itself, to electrophoresis (Fig. with rabbit antisera to tryptophan hydroxylase, goat antise- 1). Electrophoresis of individual bands, A, B, or C, resulted rum to rabbit IgG, and peroxidase-antiperoxidase complex, in the reappearance of all three bands, although in each case and then reacted for 15 min with 3,3'-,diaminobenzidine and the original band predominated (Fig. 1). This finding hydrogen proxide to yield a brown reaction product. For suggests that the inactive band B is related in some manner light microscopy the sections were lightly counterstained to the active enzyme; the reason for its apparent inactivity with cresyl violet, dehydrated, and mounted in Permount. may be due to the absence of an active prosthetic group or For electron microscopy, an area of the raphe nucleus was to insensitivity of our enzyme assay. dissected from each section, fixed for 30 min in 2% osmium tetroxide, dehydrated, and embedded in Epon 812. Ultra- Antibodies to tryptophan hydroxylase thin sections were taken from the outer surface (1-2 ,um) of We next prepared antibodies to tryptophan hydroxylase each 20-Am section for subsequent examination with a Phil- using the active enzyme of the C band (Fig. 1) as antigen. ips 301 electron microscope without additional staining. The antibody appeared highly specific for tryptophan hy- Control sections were incubated with either preimmune droxylase by the following criteria: (a) Immunoelectrophor- serum or tryptophan hydroxylase antisera maximally ad- esis and double immunodiffusion of antibody run against a sorbed with the purified enzyme (13, 15). crude extract of the raphe nuclei region of rat brain (Fig. 2, H), purified enzyme eluted from the Sephadex G-200 col- RESULTS umn (Fig. 2, G200), or to each of the specific protein bands A, B, or C (Fig. 2, A, B, and C), detected by gel electropho- Purification of tryptophan hydroxylase in rat brain resis of the Sephadex eluate (Fig. 1) yielded single precipitin Raphe nuclei from 260 Sprague-Dawley rats were homoge- arcs (Fig. 2). On the other hand, no precipitin was formed nized in 50 mM Tris-acetate buffer, pH 7.6, containing 2 when the antibody was run against other purified enzymes Downloaded by guest on October 3, 2021 Cell Biology: joh et al. Proc. Nat. Acad. Sci. USA 72 (1975) 3577

Aband- FIG. 2. Immuno double diffusion of specific antibody to C B band band of tryptophan hydroxylase (center well) against crude ho- mogenate of raphe nucleus (H), purified enzyme of Sephadex G-200 eluate (G200), and bands A, B, and C (A, B, and C, respec- tively) produced by electrophoresis, as in legend of Fig. 1. Note C'band that single precipitin arcs formed between antibody and all forms of enzyme. In sections through the midbrain raphe area near the me- dial longitudinal fasciculus, dark reaction product was con- tained in the cytoplasm of small, highly branched neurons (Fig. 3A). In perikarya, enzyme was localized in a thin rim of cytoplasm surrounding an unstained nucleus and extend- FIG. 1. Patterns on polyacrylamide gels produced by electro- ed into branched dendrites and axons which could be distin- phoresis of tryptophan hydroxylase. The gel to the left represents guished in the unstained neuropil as black dots in cross sec- the pattern produced from the G200 eluate. Three bands are iden- tion, or, when cut longitudinally and examined with phase tified: Bands A and C are enzymatically active (* represents en- contrast optics, as irregular strands of black material. The zyme activity); band B and other minor bands are enzymatically stained axons were visualized along their passage through inactive. The three remaining gels represent the pattern produced brainstem and forebrain into terminal fields. The by eluting bands A, B, and C from the G200 gel and resubmitting the upper for electrophoresis. Note that after electrophoresis of single bands, cells and processes of stained neurons were often closely as- all three original bands are detectable, although the original band sociated with heavily myelinated tracts such as the medial is preponderant. longitudinal fasciculus and medial lemniscus, and also the cortico-spinal (pyramidal) pathway. No reaction product was seen in neurons of comparable size and location in sec- involved in the biosynthesis of monoamines from rat brain, tions incubated with control serum and lightly counter- including tyrosine hydroxylase, dopamine-fl-hydroxylase, stained with cresyl violet (Fig. SB). aromatic L-amino-acid decarboxylase, and quinonoid dihy- By electron microscopy, examination of neurons in the dropterin reductase. (b) The antibody completely inhibited nucleus raphe dorsalis incubated with antibody to trypto- the activity of tryptophan hydroxylase either as purified en- phan hydroxylase demonstrated electron-dense reaction zyme (bands A and C) or in homogenates of the raphe nuclei neurons. of the other afore- product in the processes and perikarya of stained region of rat brain, but not the activities At low magnification (Fig. 3C) numerous stained unmyeli- mentioned enzymes. (c) As detailed below, the antibody was nated processes were clearly delineated from an unstained specifically localized immunohistochemically only to the cell In the cell soma bodies and processes of neurons defined by others by his- neuropil rich in myelinated fibers (Fig. 3C). tofluorescence (4, 5) as containing 5HT. (Fig. 4A) the peroxidase staining was widely distributed the other active throughout the cytoplasm. Endoplasmic reticulum and Antibodies were also prepared against Golgi apparatus were also stained, whereas mitochondria, ly- band, A, and the inactive band B. These antibodies cross- Within pro- reacted with the A, B, and C bands by immunodiffusion and sosomes, and the plasma membranes were not. electrophoresis, and inhibited enzyme activity in tissue, and cesses, the enzyme appeared to be associated with specific of the midbrain im- subcellular organelles. In longitudinal sections through both were localized only to 5HT neurons by axons and dendrites (Fig. 3C), dense reaction product was munohistochemistry. aligned in linear aggregates forming fiber-like structures ar- Immunohistochemical localization of tryptophan ranged in parallel with the plasma membrane. In cross sec- hydroxylase tion (Figs. SC, 4A, and 4B), the reaction product appeared In sections of midbrain, pons, and medulla reacted with in the form of darkly stained round structures with a mean antibody to tryptophan hydroxylase and immunohistochem- diameter of 22 nm. In small axons (Fig. 4A) the labeled sub- ically stained by the peroxidase-antiperoxidase method, a cellular organelles appeared to have a more regular geomet- dark reaction product was observed by light microscopy ric array than in dendrites. On the basis of distribution and within specific groups of neurons and their processes. The size, the labeled structures have the characteristics of mi- localization of these neurons corresponded in large measure crotubules (24) and indicate that tryptophan hydroxylase, to those areas in which 5HT neurons have been demon- like tyrosine hydroxylase (15), is in some manner associated strated by specific histofluorescence (4, 5), particularly the with that organelle. nuclei of the raphe. A detailed mapping of the location of the cell bodies and projections of processes containing tryp- DISCUSSION tophan hydroxylase will be presented elsewhere, consider- In the present study we have been able to isolate the enzyme ation being focused here only on the morphology of typical tryptophan hydroxylase from rat brain of sufficient purity so cells. as to produce antibodies to the enzyme. The principal obsta- Downloaded by guest on October 3, 2021 3578 Cell Biology: joh et al. Proc. Nat. Acad. Sci. USA 72 (1975)

. >; '~ f FIG. 3. Immunohistochemical localization of tryptophan hy- FIG. 4. Ultrastructural localization of tryptophan hydroxylase droxylase to neurons in the region of the raphe nuclei of rat brain in soma and processes of neurons in nucleus raphe dorsalis using mesencephalon stained by the peroxidase-antiperoxidase method. peroxidase-antiperoxidase staining method. (A) Electron micro- (A) Phase contrast photomicrograph of section incubated with graph of section incubated with specific tryptophan hydroxylase tryptophan hydroxylase antiserum showing darkly stained reac- antiserum showing staining in cytoplasm of a neuron whose lateral tion product in the cytoplasm and processes of neurons (unmarked border is indicated by arrows. There is some selective association arrows). Other labeled neuronal processes (P) are seen in close as- with endoplasmic reticulum (ER), but not with mitochondria (M). sociation with unstained myelinated fiber bundles (MB) X132.5. Axon (Ax) in neuropil shows selective staining of rounded subunits a 22-24 nm in diameter.~BX21,200. (B) Higher magnification photo- (B) Control phase contrast photomicrograph of section through the raphe nuclei incubated with blocked antiserum. Unstained micrograph of dendrite indicated by the synapse (S) in Fig. 3C. neurons (arrows) are associated with the myelinated fiber bundle Numerous round peroxidase-stained structures with dimensions of (MB) X132.5. (C) Low magnification electron micrograph of per- neurotubules (arrows) are present throughout the dendrite indi- oxidase staining in neuronal processes of section incubated with cated by synapse (S). Mitochondria (M) are unstained. X21,200. tryptophan hydroxylase antiserum. A longitudinal section through (C) Control electron micrograph of a section incubated with one dendritic process shows selective localization along neurotubu- blocked antiserum. The dendrite whose surface is marked by the lar structures (arrows). A cross section through a labeled dendrite synapse (S) on its surface contains unstained neurotubules (ar- is indicated by the synapse (S) on its surface. Cross sections of nu- rows) as well as unstained mitochondria (M). X21,200. Darkening merous other labeled processes are distributed throughout the sec- of the cell membrane in the area of the synaptic cleft occurs in tion. Unlabeled myelinated axons (MA) are also present X10,600. both control and stained sections and hence is not an indication that tryptophan hydroxylase is localized to the synapse. cle to purification of tryptophan hydroxylase, which has permitted only partial purification by others (9, 10), is the slowly moving A band, and the inactive band B, and that tendency of the enzme to aggregate into polymers of high each band when isolated and resubjected to electrophoresis molecular weight. Such aggregation results in a low yield of resulted in replication of the three bands suggests that tryp- enzyme activity with little increase in specific activity. It tophan hydroxylase exists in multiple forms, as others have also reduces the mobility of enzyme by electrophoresis, suggested (3, 8). The nature of the multiple forms of the en- thereby impairing its separation from other proteins which zyme, however, is uncertain. The fact that antibodies to one is essential for production of specific antibodies. We were band crossreact with each other and that active bands A and able to eliminate aggregation by passage of the enzyme C were inhibited by p-chlorophenylalanine suggests that through a Sepharose 4B column. This step removed in the each form of the enzyme shares a common protein structure, void volume enzymatically inactive substances of high mo- and that at least bands A and C have closely related active lecular weight containing a high content of nucleic acids sites. However, whether the multiple forms of the enzyme whose presence appears to facilitate aggregation of trypto- represent different states of aggregation of basic enzyme phan hydroxylase. Comparable macromolecules associated units into dimers, tetramers, etc., as is the case with tyrosine with nucleotides also appear to facilitate aggregation of ty- hydroxylase (12, 25), or represent charge isomers, i.e., forms rosine hydroxylase (12). of the enzyme having comparable protein structure but dif- The observations that electrophoresis of highly purified fering in ionic charge, as exemplified by phenylethanolam- tryptophan hydroxylase consistently produced three major ine-N-methyl transferase (26), remains to be determined. protein bands, a highly active mobile C band, a less active, By immunohistochemistry, tryptophan hydroxylase was Downloaded by guest on October 3, 2021 Cell Biology: Joh et al. Proc. Nat. Acad. Sci. USA 72 (1975) 3579

localized by light microscopy to the cytoplasm of small peu- central nervous system raises for the first time the possibility rons with extensively branched processes, many of which are that this enzyme is also transported, conceivably by mecha- concentrated within the confines of the nuclei of the raphe. nisms similar to those for tyrosine hydroxylase. These neurons, on the basis of their configuration and distri- We thank Marcia Brodsky and Nancy Shieh for excellent techni- bution, correspond almost exactly to those defined by his- cal help. The research was supported by grants from the National tofluorescence as containing the neurotransmitter 5HT (4, Institutes of Health (MH 24285, NS 06911) and the Benevolent 5). These findings would support the indirect evidence that Foundation of Scottish Rite Freemasonry. in brain, tryptophan hydroxylase is contained exclusively within those neurons that synthesize, store, and release the 1. Koe, K. & Weissman, A. (1966) J. Pharmacol. Exp. Ther. 154, neurotransmitter 5HT. 499-516. Ultrastructurally the distribution of tryptophan hydroxyl- 2. Jequier, E., Lovenberg, W. & Sjoerdsma, A. (1967) Mol. Phar- ase is not uniform throughout the cell. Within the cell body macol. 3,274-278. & A. Sci. 761-772. the enzyme is diffusely distributed in cytoplasm but tends to 3. Knapp, S. Mandell, J. (1972) Life 11, membranes of reticulum 4. Dahlstrom, A. & Fuxe, K. (1964) Acta Physiol. Scand. 62, be associated with endoplasmic suppl. 232, 1-55. and Golgi apparatus, possibly reflecting the sites in which 5. Fuxe, K. & Johnson, G. (1967) Histochemie 11, 161-166. the specific protein is synthesized (27). In processes, both 6. Jequier, E., Robinson, D. S., Lovenberg, W. & Sjoerdsma, A. axons and dendrites, the enzyme is primarily associated with (1969) Biochem. Pharmacol. 18, 1071-1081. subcellular organelles having the characteristics of microtub- 7. Gal, E. M., Roggeveen, A. E. & Millard, S. A. (1970) J. neuro- ules (24). chem. 17, 1224-1235. The nature of the association of tryptophan hydroxylase 8. Gal, E. M. (1974) in Advances in Biochemical Psychophar- and microtubules in axons and dendrites is obscure. Con- macology, eds. Costa, E., Gessa, G. L. & Sandler, M. (Raven ceivably it might be an artifact in which the enzyme, which Press, New York), Vol. 11, pp. 1-11. is preponderantly soluble (3, 8), becomes cross-linked to 9. Friedman, P. A., Kappelman, A. H. & Kaufman, S. (1972) J. Biol. Chem. 247, 4165-4173. as a consequence glu- other cellular proteins of fixation with 10. Ichiyama, A., Nakamura, S., Nishizuka, Yl. & Hayaishi, 0. taraldehyde. However, for cross-linking to occur the enzyme (1970) J. Biol. Chem. 245, 1699-1709. and the cellular protein to which it binds must be in close 11. Joh, T. H., Geghman, C. & Reis, D. J. (1973) Proc. Nat. Acad. proximity since free diffusion would be arrested by fixation. Sci. USA 70,2767-2771. Thus the fact that only microtubules are stained for trypto- 12. Joh, T. H. & Reis, D. J. (1975) Brain Res. 85, 146-151. phan hydroxylase while other membranous constituents of 13. Pickel, V. M., Joh, T. H., Field, P. M., Becker, C. G. & Reis, the cell, such as lysosomes, mitochondria, or plasma mem- D. J. (1975) J. Histochem. Cytochem. 23, 1-12. branes, are not, indicates that the association of enzyme and 14. Pickel, V. M., Joh, T. H. & Reis, D. J. (1975) Brain Res. 85, microtubules has specificity. It is not possible to determine 295-300. by these methods whether tryptophan hydroxylase is incor- 15. Pickel, V. M., Joh, T. H. & Reis, D. J. (1975) Proc. Nat. Acad. Sci. USA 72,659-663. porated into microtubules, is adherent to their surface, or is 16. Deguchi, T., Shinka, A. K. & Barchas, J. D. (1973) J. Neuro- conveyed free in cytosol but in close approximation to the chem. 20, 1329-1336. organelle. 17. Christenson, J. G., Dairman, W. & Udenfriend, S. (1970) The ultrastructural distribution of tryptophan hydroxyl- Arch. Biochem. Biophys. 141, 356-363. ase, and in particular its localization to microtubules, is strik- 18. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. ingly similar to that of the closely related enzyme tyrosine (1951) J. Biol. Chem. 193,265-275. hydroxylase (14, 15), which catalyzes the rate-limiting step 19. Davis, B. J. (1964) Ann. N.Y. Acad. Sci. 121, 404-427. in the biosynthesis of catecholamines in the sympathetic ner- 20. Ornstein, L. (1964) Ann. N.Y. Acad. Sci. 121,321-349. vous system and in brain. The association of both tyrosine 21. Ouchterlony, 0. (1958) in Progress in Allergy, ed. Kallos, P. and tryptophan hydroxylases with microtubules is of consid- (Karger, Basel), Vol. V, pp. 1-78. 22. Scheidegger, J. J. (1955) Int. Arch. Allergy Appl. Immunol. 7, erable interest with regard to the potential transport of these 103-110. enzymes from the presumed sites of synthesis in the cell 23. Sternberger, L. A., Hardy, P. H., Cuculis, J. J. & Myer, H. G. bodies into terminals within the brain. The fact that in pe- (1970) J. Histochem. Cytochem. 18, 315-333. ripheral sympathetic neurons tyrosine hydroxylase is trans- 24. Wuerker, R. B. & Kirkpatrick, J. B. (1972) Int. Rev. Cytol. 33, ported at relatively rapid rates of flow (27, 28) and that the 45-76. integrity of microtubules appears important in mediating 25. Musacchio, J. M., Wurzberger, R. J. & D'Angelo, G. L. (1971) fast transport (29, 30) has led us to suggest that the associa- Mol. Pharmacol. 7, 136-146. tion of tyrosine hydroxylase with microtubules in the central 26. Joh, T. H. & Goldstein, M. (1973) Mol. Pharmacol. 9, 117- nervous system provides indirect evidence for the transport 129. 27. Geffen, L. B. & Livett, B. G. (1971) Physiol. Rev. 517,98-157. enzyme On of this within the brain (14, 15). the other hand, 28. Jarrot, B. & Geffen, L. B. (1972) Proc. Nat. Acad. Sci. USA it is not known whether tryptophan hydroxylase is also trans- 69,3440-3442. ported from the cell body into axons to the periphery of 29. Thoenen, H., Otten, U. & Oesch, F. (1973) Brain Res. 62, 5HT neurons, because no peripheral model of this neuron 471-475. exists. However, the present finding of an association of 30. Banks, P., Mayor, D. & Tomlinson, D. R. (1971) J. Physiol. tryptophan hydroxylase with the microtubules in the rat (London) 219,755-761. Downloaded by guest on October 3, 2021