Amphetamine (AMPH) and its derivatives embody a remarkable breadth of pharmacology. These molecules exert their effects, both therapeutic and pathological, at the human monoamine transporters, which tune synaptic dynamics by evacuating monoamine neuromodulators from the synapse subsequent to neuronal impulses. These transporters are electrogenic, and the transporter-mediated current can be correlated to a surrogate measure of the change in membrane voltage: Ca++ currents from co-transfected L-type Ca++ channels. The present work makes use of this assay, with which it is possible to derive pharmacodynamic metrics from both substrates and inhibitors. This work presents data on a heretofore-unstudied class of amphetamine analogs: the enantiomers of N-Me 4-Me AMPH and N-Et 4-Me AMPH. Remarkably, while both enantiomers of the N-Me version of this compound function as substrates at hDAT, both enantiomers of the N-Et version are inhibitors. This switch does not occur at hNET, where all enantiomers of both N-Me and N-Et 4-Me AMPH function as substrates. Further, (S)-N-Et 4-Me AMPH is a substrate at dDAT. EC50 and IC50 values for all drugs at both transporters are presented. I present the results of super-resolution microscopic co-localization studies on the plasmalemmal spatial relation of the human dopamine transporter and voltage gated calcium channel, L-type 1.2 (CaV1.2). I discuss future aims toward a unified understanding of the mechanisms of monoamine transporter function, with an emphasis on what amphetamine can illuminate in this regard.

O peptídeo idêntico ao C-terminal da proteína S100A9 murina (pS100A9mH92-G110) inibe a hiperalgesia inflamatória induzida pela carragenina. Em adição, este peptídeo inibe a hiperalgesia inflamatória induzida por tripsina, uma serino protease capaz de ativar receptores ativados por protease do tipo 2 (PAR2). O objetivo inicial deste trabalho foi caracterizar a relação estrutura/ efeito do pS100A9m, a fim de determinar a menor seqüência peptídica dotada de atividade antinociceptiva. Ainda, como parte dos objetivos, neste trabalho foram investigados os mecanismos envolvidos no efeito antinociceptivo do pS100A9m e da menor seqüência ativa sobre a hiperalgesia induzida pela ativação de PAR2. Diferentes seqüências peptídicas homólogas ao pS100A9m foram sintetizadas e avaliadas em ratos submetidos ao modelo de hiperalgesia mecânica induzida por carragenina. Dentre todas as seqüências peptídicas investigadas, o peptídeo denominado AcE97-G102 foi determinado como a menor seqüência ativa com efeito semelhante ao pS100A9m. Com relação aos estudos sobre a ativação de PAR2, os resultados obtidos demonstraram que o pS100A9m bem como o AcE97-G102 inibem a hiperalgesia térmica e mecânica decorrentes da ativação de PAR2 (induzida por um peptídeo agonista deste receptor ? PAR2AP). A análise por imuno-histoquímica demonstrou que a ativação de PAR2 aumenta a expressão da proteína Egr-1 em neurônios nociceptivos, sendo o pS100A9m capaz de inibir este efeito. Em adição, ambos pS100A9m e AcE97-G102 inibiram o influxo de cálcio induzido por PAR2AP ou tripsina, em neurônios sensoriais do gânglio da raiz dorsal da medula espinhal (DRG). Por outro lado, nenhum dos peptídeos apresentou efeito sobre a mobilização de cálcio em células HEK-293, que naturalmente expressam PAR2, ou em células KNRK transfectadas com este tipo de receptor, sugerindo que o efeito tanto do pS100A9m quanto do AcE97-G102, sobre a ativação de PAR2, seja específico para neurônios sensoriais. O pS100A9m e o AcE97-G102 inibiram o influxo de cálcio nos neurônios DRG estimulados com bradicinina, capsaicina ou KCl. Ainda, o pS100A9m inibiu a liberação de substância P induzida por PAR2. Os resultados obtidos com o tratamento de neurônios DRG com tapsigaragina ou com ionóforo de cálcio sugerem um efeito direto do pS100A9m sobre os canais de cálcio. Desta forma, foi avaliada atividade do pS100A9m e do AcE97-G102 sobre culturas de células HEK-tsA transfectadas com canais de cálcio dependente de voltagem do tipo N ou do tipo L. Os resultados obtidos demonstraram que ambos peptídeos inibirem o influxo de cálcio em células transfectadas com receptores do tipo N. Em conjunto, os dados aqui obtidos demonstram que o efeito do C-terminal da proteína S100A9 murina sobre a nocicepção experimental é devido a uma inibição de canais de cálcio do tipo N, por uma ação direta em neurônios sensoriais. Ainda, a seqüência responsável por este efeito está localizada na porção E97-G102 do domínio C-terminal da proteína S100A9 murina. / Peptide identical to the C-terminus of S100A9 protein (mS100A9pH92-G110) inhibits inflammatory hyperalgesia induced by carrageenan and trypsin, a serine protease that activates protease-activated receptors 2 (PAR2). The aim of this work was to characterize the relationship between structure and function of mS100A9p in order to identify the shortest peptide sequence endowed with antinociceptive effect. Furthermore, the mechanisms involved on the antinociceptive effect of both mS100A9p and the shortest homologous sequence on PAR2-induced hyperalgesia were also evaluated. Different peptide sequences homologous to mS100A9p were synthesized and evaluated in rats submitted to the carrageenan-induced mechanical hyperalgesia model. Among all evaluated sequences, the peptide AcE97-G102 was found to be the shortest sequence that showed an antinociceptive effect similar to that induced by mS100A9p. In regard to PAR2 activation, data obtained herein demonstrated that both mS100A9p and AcE97-G102 inhibit PAR2-induced mechanical and thermal hyperalgesia, induced by the selective agonist peptide ? PAR2AP. Imunohistochemical evaluation demonstrated that PAR2 activation increased Egr-1 protein expression on sensory neurons and mS100A9p inhibited this effect. In addition, both mS100A9p and AcE97-G102 inhibited PAR2- and trypsin-induced calcium influx in dorsal root ganglia neurons (DRG). On the other hand, no effect on the calcium influx of the peptides were observed on HEK-293 cells or KNRK-PAR2 transfected cells, suggesting that the effects of mS100A9p and AcE97-G102 on PAR2 activation are specific for sensory neurons. Both mS100A9p and AcE97-G102 inhibited DRG calcium flux when cells were stimulated with bradykinin, capsaicin or KCl. Also, mS100A9p inhibited PAR2-induced substance P release in DRG. Treatment of DRG with either thapsigargin or calcium ionophore suggest a direct effect of mS100A9p on calcium channels. To evaluate this hypothesis the effects of mS100A9p and AcE97-G102 were evaluated on N-type or L-type voltage-dependent calcium channel transfected HEK-tsA cells. Both peptides inhibited calcium influx of N-type transfected cells. In conclusion, data presented herein demonstrate that the C-terminus of murine S100A9 protein inhibits experimental nociception through a block of N-type voltage-dependent calcium channels, directly on sensory neurons. Also, the domain involved in this effect is localized on the sequence E97-G102 of the C-terminus of murine S100A9 protein.

Bien que la régulation des canaux calciques par les récepteurs couplés aux protéines G soit connue depuis une trentaine d'année, ce n'est que récemment qu'il a été découvert que ce phénomène pouvait passer par une interaction directe entre ces deux partenaires. Les RCPGs sont les senseurs d'un grand nombre de paramètres (des simples photons aux molécules odorantes en passant par des hormones, acides aminés et nucléotides) et ils contrôlent un grand nombre de processus cellulaires en fonction de ces différents stimuli, ce qui en fait une cible thérapeutique majeure. Une de leurs cibles est l'activité des canaux calciques voltage dépendants qui est responsable d'un grand nombre de processus tels que le contrôle du potentiel de membrane, le relargage de neurotransmetteurs, la contraction musculaire ou, bien sûr, le contrôle du taux de calcium intracellulaire qui est lui-même un second messager impliqué dans de nombreuses voies de régulations.Il nous a donc paru intéressant de se pencher plus en avant sur ces interactions et de trouver une méthode nous permettant de cribler ces interactions potentielles avec des RCPG ciblés pouvant intervenir dans une thématique de contrôle de la douleur. Pour cela nous avons développé une stratégie de FRET en temps résolu utilisant les cryptates de terres rares à déactivation lente couplés aux ligands du tag SNAP comme donneurs de fluorescence, les canaux calciques étudiées étant fusionnés avec cet épitope et l'eGFP fusionnée aux RCPGs en tant qu'accepteur. Ce test nous a permis de confirmer l'interaction entre CaV2.2 et ORL1 le récepteur de la nociceptine. Nous avons ensuite cherché à caractériser plus précisément cette interaction et nous avons déterminé quelles en étaient les séquences peptidiques responsables au sein des domaines C-terminaux de ces deux protéines grâce à des expériences de GST-pull down. Nous avons synthétisé un peptide reproduisant la séquence d'interaction d'ORL1 que nous avons couplé à la séquence TAT, le rendant ainsi capable de pénétrer les membranes cellulaires. Lorsque nous ajoutons ce peptide leurre dans les expériences de TR-FRET, l'augmentation de fluorescence observée en présence de CaV2.2-SNAP et ORL1-GFP disparait totalement alors que l'ajout d'un peptide contrôle composé des mêmes acides aminés mais présentés dans le désordre n'a aucun effet. Nous avons ensuite cherché à étudier les effets de ce peptide in vivo lors d'un protocole de tolérance à la morphine étant donné que les souris K.O. pour le gène d'ORL1 sont résistantes à l'apparition de cette tolérance. Cette stratégie de découplage CaV2.2 :: ORL1 abolit complètement le phénomène de tolérance aux effets analgésiques de la morphine par une action au niveau spinal. Ce travail peut conduire à l'utilisation d'une telle approche dans une perspective thérapeutique visant à améliorer l'utilisation de morphiniques lors du traitement des douleurs chroniques. / The regulation of the calcium channels by GPCRs has been known for almost thirty years but the direction interaction between those two proteins is a recent breakthrough. GPCRs are sensors for a great number of parameters (photons, smell molecules, hormones, amino acids, nucleotides…) and they control numerous cellular functions according to those parameters making them a major target for pharmacology. One of the GPCR's targets is the calcium channel activity which is responsible for a great number of cellular processes like control of the membrane potential, neurotransmitters or hormonal secretion, muscular contraction and, of course, control of the intracellular calcium level which is a second messenger of numerous cell-regulation pathways.It appears to us that it would be interesting to study more closely those interactions and find a way to screen the GPCR/calcium channels interactions that may occur in pain regulation. We developed a strategy of time resolved FRET, using rare earth cryptate coupled to the ligand of the SNAP tag which is fused to the calcium channel as fluorescence donor and eGFP fused GRPRs as acceptors. That test confirmed the interaction between CaV2.2 and ORL1, the nociceptin receptor. We characterized more precisely the peptide sequence of the carboxy-terminal domain of the two proteins which is responsible for the interaction using GST-pull down experiments. We synthesized a peptide reproducing the ORL1 interaction sequence coupled to the TAT sequence allowing to go through the cell membranes. When we add this decoy peptide to ours TR-FRET experiments we lose all the increase of fluorescence that we see in presence of CaV2.2-SNAP and ORL1-GFP but the adding of a control peptide made of the same peptides but scrambled didn't affect the experiment. Then we look for the effects of this peptide in vivo, during a morphine tolerance protocol as it was reported that the ORL1 knock-out mice were insensitive to this phenomenon. This strategy of uncoupling CaV2.2 and ORL1 leads to a complete suppression of the tolerance to the analgesic effects of the morphine by an action at the spinal level. This work could lead to a therapeutic use of this approach which could enhance the use of morphinic compounds in treatment of chronics pains.

O peptídeo idêntico ao C-terminal da proteína S100A9 murina (pS100A9mH92-G110) inibe a hiperalgesia inflamatória induzida pela carragenina. Em adição, este peptídeo inibe a hiperalgesia inflamatória induzida por tripsina, uma serino protease capaz de ativar receptores ativados por protease do tipo 2 (PAR2). O objetivo inicial deste trabalho foi caracterizar a relação estrutura/ efeito do pS100A9m, a fim de determinar a menor seqüência peptídica dotada de atividade antinociceptiva. Ainda, como parte dos objetivos, neste trabalho foram investigados os mecanismos envolvidos no efeito antinociceptivo do pS100A9m e da menor seqüência ativa sobre a hiperalgesia induzida pela ativação de PAR2. Diferentes seqüências peptídicas homólogas ao pS100A9m foram sintetizadas e avaliadas em ratos submetidos ao modelo de hiperalgesia mecânica induzida por carragenina. Dentre todas as seqüências peptídicas investigadas, o peptídeo denominado AcE97-G102 foi determinado como a menor seqüência ativa com efeito semelhante ao pS100A9m. Com relação aos estudos sobre a ativação de PAR2, os resultados obtidos demonstraram que o pS100A9m bem como o AcE97-G102 inibem a hiperalgesia térmica e mecânica decorrentes da ativação de PAR2 (induzida por um peptídeo agonista deste receptor ? PAR2AP). A análise por imuno-histoquímica demonstrou que a ativação de PAR2 aumenta a expressão da proteína Egr-1 em neurônios nociceptivos, sendo o pS100A9m capaz de inibir este efeito. Em adição, ambos pS100A9m e AcE97-G102 inibiram o influxo de cálcio induzido por PAR2AP ou tripsina, em neurônios sensoriais do gânglio da raiz dorsal da medula espinhal (DRG). Por outro lado, nenhum dos peptídeos apresentou efeito sobre a mobilização de cálcio em células HEK-293, que naturalmente expressam PAR2, ou em células KNRK transfectadas com este tipo de receptor, sugerindo que o efeito tanto do pS100A9m quanto do AcE97-G102, sobre a ativação de PAR2, seja específico para neurônios sensoriais. O pS100A9m e o AcE97-G102 inibiram o influxo de cálcio nos neurônios DRG estimulados com bradicinina, capsaicina ou KCl. Ainda, o pS100A9m inibiu a liberação de substância P induzida por PAR2. Os resultados obtidos com o tratamento de neurônios DRG com tapsigaragina ou com ionóforo de cálcio sugerem um efeito direto do pS100A9m sobre os canais de cálcio. Desta forma, foi avaliada atividade do pS100A9m e do AcE97-G102 sobre culturas de células HEK-tsA transfectadas com canais de cálcio dependente de voltagem do tipo N ou do tipo L. Os resultados obtidos demonstraram que ambos peptídeos inibirem o influxo de cálcio em células transfectadas com receptores do tipo N. Em conjunto, os dados aqui obtidos demonstram que o efeito do C-terminal da proteína S100A9 murina sobre a nocicepção experimental é devido a uma inibição de canais de cálcio do tipo N, por uma ação direta em neurônios sensoriais. Ainda, a seqüência responsável por este efeito está localizada na porção E97-G102 do domínio C-terminal da proteína S100A9 murina. / Peptide identical to the C-terminus of S100A9 protein (mS100A9pH92-G110) inhibits inflammatory hyperalgesia induced by carrageenan and trypsin, a serine protease that activates protease-activated receptors 2 (PAR2). The aim of this work was to characterize the relationship between structure and function of mS100A9p in order to identify the shortest peptide sequence endowed with antinociceptive effect. Furthermore, the mechanisms involved on the antinociceptive effect of both mS100A9p and the shortest homologous sequence on PAR2-induced hyperalgesia were also evaluated. Different peptide sequences homologous to mS100A9p were synthesized and evaluated in rats submitted to the carrageenan-induced mechanical hyperalgesia model. Among all evaluated sequences, the peptide AcE97-G102 was found to be the shortest sequence that showed an antinociceptive effect similar to that induced by mS100A9p. In regard to PAR2 activation, data obtained herein demonstrated that both mS100A9p and AcE97-G102 inhibit PAR2-induced mechanical and thermal hyperalgesia, induced by the selective agonist peptide ? PAR2AP. Imunohistochemical evaluation demonstrated that PAR2 activation increased Egr-1 protein expression on sensory neurons and mS100A9p inhibited this effect. In addition, both mS100A9p and AcE97-G102 inhibited PAR2- and trypsin-induced calcium influx in dorsal root ganglia neurons (DRG). On the other hand, no effect on the calcium influx of the peptides were observed on HEK-293 cells or KNRK-PAR2 transfected cells, suggesting that the effects of mS100A9p and AcE97-G102 on PAR2 activation are specific for sensory neurons. Both mS100A9p and AcE97-G102 inhibited DRG calcium flux when cells were stimulated with bradykinin, capsaicin or KCl. Also, mS100A9p inhibited PAR2-induced substance P release in DRG. Treatment of DRG with either thapsigargin or calcium ionophore suggest a direct effect of mS100A9p on calcium channels. To evaluate this hypothesis the effects of mS100A9p and AcE97-G102 were evaluated on N-type or L-type voltage-dependent calcium channel transfected HEK-tsA cells. Both peptides inhibited calcium influx of N-type transfected cells. In conclusion, data presented herein demonstrate that the C-terminus of murine S100A9 protein inhibits experimental nociception through a block of N-type voltage-dependent calcium channels, directly on sensory neurons. Also, the domain involved in this effect is localized on the sequence E97-G102 of the C-terminus of murine S100A9 protein.

The auditory brainstem is comprised of a number of synapses specialized for the transmission of high-fidelity synaptic signals. Within the first three postnatal weeks, these pathways acquire the ability to process high-frequency signals without compromising timing information. However, little is known regarding developmental adaptations which confer this ability. Situated in the sound localization pathway, the calyx of Held-medial nucleus of the trapezoid body synapse provides an ideal model for investigating such adaptations as both the pre- and postsynaptic neurons are accessible to electrophysiological experimentation. Using this synapse, we have shown herein that the spatial coupling between voltage-gated calcium channels (VGCCs) and synaptic vesicles (SVs) tightens during development. Immature synapses use a loosely-coupled arrangement of many N- and P/Q-type VGCCs (“microdomain” modality) while mature synapses use a tightly-coupled arrangement of fewer P/Q-type VGCCs, to release SVs (“nanodomain” modality). As a consequence of this tightening, synaptic delay (SD) shortens. By fluorescence- and electron microscopy of SVs near active zones, we further identified the filamentous protein septin 5 as a molecular substrate, differentiating the two release modalities, which may act as a spatial barrier separating VGCCs and SVs in immature synapses. Finally, we have demonstrated that changes in release modality affect the nature of short-term plasticity observed at this synapse. Using trains of action potentials as presynaptic voltage-commands, we showed that, downstream of calcium influx, the microdomain modality promotes short-term facilitation in excitatory postsynaptic currents (IEPSC), and calcium-dependent decreases in SD, with these being absent in synapses employing the nanodomain modality. In contrast, we found that as a result of depletion of SVs, short-term depression of IEPSC dominates in synapses using the nanodomain modality, and correlates with calcium-dependent increases in SD. These findings imply that the type of release modality has a significant impact on the strength and timing of synaptic responses. The microdomain modality imparts greater dynamic range in timing and strength, but does so at the cost of efficiency and fidelity, while the nanodomain modality is a key accomplishment consolidating the high-fidelity abilities of this synapse.

The auditory brainstem is comprised of a number of synapses specialized for the transmission of high-fidelity synaptic signals. Within the first three postnatal weeks, these pathways acquire the ability to process high-frequency signals without compromising timing information. However, little is known regarding developmental adaptations which confer this ability. Situated in the sound localization pathway, the calyx of Held-medial nucleus of the trapezoid body synapse provides an ideal model for investigating such adaptations as both the pre- and postsynaptic neurons are accessible to electrophysiological experimentation. Using this synapse, we have shown herein that the spatial coupling between voltage-gated calcium channels (VGCCs) and synaptic vesicles (SVs) tightens during development. Immature synapses use a loosely-coupled arrangement of many N- and P/Q-type VGCCs (“microdomain” modality) while mature synapses use a tightly-coupled arrangement of fewer P/Q-type VGCCs, to release SVs (“nanodomain” modality). As a consequence of this tightening, synaptic delay (SD) shortens. By fluorescence- and electron microscopy of SVs near active zones, we further identified the filamentous protein septin 5 as a molecular substrate, differentiating the two release modalities, which may act as a spatial barrier separating VGCCs and SVs in immature synapses. Finally, we have demonstrated that changes in release modality affect the nature of short-term plasticity observed at this synapse. Using trains of action potentials as presynaptic voltage-commands, we showed that, downstream of calcium influx, the microdomain modality promotes short-term facilitation in excitatory postsynaptic currents (IEPSC), and calcium-dependent decreases in SD, with these being absent in synapses employing the nanodomain modality. In contrast, we found that as a result of depletion of SVs, short-term depression of IEPSC dominates in synapses using the nanodomain modality, and correlates with calcium-dependent increases in SD. These findings imply that the type of release modality has a significant impact on the strength and timing of synaptic responses. The microdomain modality imparts greater dynamic range in timing and strength, but does so at the cost of efficiency and fidelity, while the nanodomain modality is a key accomplishment consolidating the high-fidelity abilities of this synapse.

Activity of the Ryanodine Receptor (RyR2) (aka cardiac Ca2+ release channel) plays a pivotal role in contraction of the heart. S-adenosyl-l-methionine (SAM) is a biological methyl group donor that has close structural similarity to ATP, an important physiological regulator of RyR2. This work provides evidence that SAM can act as a RyR2 regulatory ligand in a manner independent from its recognized role as a biological methyl group donor. RyR2 activation appears to arise from the direct interaction of SAM, via its adenosyl moiety, with the RyR2 adenine nucleotide binding sites. Because uncertainty remains regarding the structural motifs involved in RyR2 modulation by ATP and its metabolites, this finding has important implications for clarifying the structural basis of ATP regulation of RyR2. During the course of this project, direct measurements of single RyR2 activity revealed that SAM has distinct effects on RyR2 conductance. From the cytosolic side of the channel, SAM produced a single clearly resolved subconductance state. The effects of SAM on channel conductance were dependent on SAM concentration and membrane holding potential. A second goal of this work was to distinguish between the two possible mechanisms by which SAM could reduce RyR2 conductance: i) SAM interfering directly with ion permeation via binding within the conduction pathway (pore block), or ii) SAM binding a regulatory (or allosteric) site thereby stabilizing or inducing a reduced conductance conformation of the channel. It was determined that SAM does not directly interact with the RyR2 conduction pathway. To account for these observations an allosteric model for the effect of SAM on RyR2 conductance is proposed. According to this model, SAM binding stabilizes an inherent RyR2 subconductance conformation. The voltage dependence of the SAM related subconductance state is accounted for by direct effects of voltage on channel conformation which indirectly alter the affinity of RyR2 for SAM. Patterns in the transitions between RyR2 conductance states in the presence of SAM may provide insight into the structure-activity relationship of RyR2 which can aid in the development of therapeutic strategies targeting this channel.

In this dissertation I introduce a simple experimental approach for studying afferent pain fibre physiology. I developed an en bloc dural-skull preparation that pairs electrophysiological stimulations, pharmacological manipulations, and the UV photolysis of caged compounds in and around selectively identified individual C-fibre nociceptors with microfluorometric imaging of Ca2+ responses. This allows the observation of physiological functioning in individual nociceptive fibre free nerve endings. I show high-resolution functional imaging of single action potential-evoked fluorescent transients, as well as sub- and supra-threshold calcium signaling events within individual nociceptive fibre terminations. Utilizing the dural-skull preparation I was able to identify a peripheral mechanism of action in the terminals of CGRP nociceptive fibres for an effective migraine therapeutic, the selective 5-HT1 receptor agonist, sumatriptan. I found sumatriptan to cause an approximately 40% reduction in the amplitude of action potential-evoked Ca2+ transients in the peripheral terminals of CGRP nociceptive fibres that was mediated selectively through the inhibition of N-type Ca2+ channels. Observations from this study support a peripheral site of action for sumatriptan in inhibiting the activity of dural pain fibres and adds to our understanding of the mechanisms that underlie the clinical effectiveness of 5-HT1 receptor agonists such as sumatriptan. While μ-opioid receptor agonists remain the most powerful analgesics for the treatment of severe pain, their mechanism of action in peripheral primary afferent pain fibres remain to be established. Further exploiting the dural-skull preparation I found activation of μ-opioid receptors in individual CGRP terminals had a dual modulatory effect; inhibition of N-type Ca2+ channel signaling and a frequency dependent, BKCa channel-mediated, suppression of action potential firing. These results establish possible anti-nociceptive mechanisms of μ-opioid receptor activation in the peripheral terminals of CGRP nociceptive fibres and identify new pathways to target for peripherally mediated analgesia. The development and subsequent testing of the dural-skull preparation in this dissertation displays its utility and opens up a new window for studying nociceptive fibre physiology and pathophysiology.

Diss. (sammanfattning) Stockholm : Karol. inst., 2004. / Härtill 4 uppsatser.

Thesis (Ph.D.)--University of South Florida, 2005. / Includes vita. Includes bibliographical references (leaves 128-144). Also available online.