J Comp Physiol A (1998) 183: 753±758
Ó Springer-Verlag 1998
ORIGINAL PAPER
F.-Y. Zhao á A. Roberts
Assessing the roles of glutamatergic and cholinergic synaptic drive in the control of ®ctive swimming frequency in young Xenopus tadpoles Accepted: 14 August 1998
Abstract This paper investigates the proposal that the frequency of the swimming central pattern generator in young Xenopus tadpoles is partly determined by the population of glutamatergic premotor interneurons active on each cycle. During ®ctive swimming spinal neurons also receive cholinergic and electrotonic excitation from motoneurons. As frequency changes during swimming we make two predictions: ®rst, since most motoneurons ®re very reliably at all frequencies, the electrotonic and nicotinic drive from motoneurons should remain constant, and second, when swimming frequency decreases, the glutamatergic drive should decrease as the number of active premotor excitatory interneurons decreases. We have tested these predictions by measuring the excitatory synaptic drive to motoneurons as frequency changes during ®ctive swimming. The components of synaptic drive were revealed by the local microperfusion of strychnine together with dierent excitatory antagonists. After blocking the nicotinic acetylcholine receptor, the mainly glutmatergic excitatory synaptic drive still changed with frequency. However, when glutamate receptors or all chemical transmission was blocked, excitation did not change with frequency. Our predictions are con®rmed, suggesting that premotor excitatory interneurons are a major factor in frequency control in the tadpole central pattern generator and that motoneurons provide a stable background excitation. Key words Tonic depolarization á EPSP á Excitatory amino acids á Locomotion á Frequency control Abbreviations ACh acetylcholine á AMPA (RS)-aamino-3-hydroxy-5-methyl-4-isoxazole propionic acid á DHbEdihyro-b-erythroidine á DMSOdimethylsulfoxide á F.-Y. Zhao á A. Roberts (&) School of Biological Sciences, University of Bristol, Bristol BS8 1UG, UK e-mail:
[email protected] Fax: +44-117-9257374
EPSP excitatory post-synaptic potential á GlutR glutamate receptor á IPSP inhibitory post-synaptic potential á nAChR nicotinic acetylcholine receptor á NBQX 6-nitro-7-sulfamoylbenzo[f]-quinoxaline-2,3dione á NMDA N-methyl-D-aspartic acid
Introduction Most rhythmic motor activities depend fundamentally on the activity of a central pattern generator in the central nervous system (see Stein et al. 1997). This paper is concerned with the mechanisms that control the frequency of central pattern generators. In many cases rhythmic movements start at a high frequency and then tend to slow down before ®nally stopping, for example scratching in turtles (Robertson et al. 1985), swimming in lampreys (McClellan and Grillner 1983) and swallowing in mammals (Jean 1984). We would like to understand what mechanisms lead to such natural changes in frequency and we have examined this problem in the swimming of young frog tadpoles. Swimming locomotion in Xenopus tadpoles is generated by a central pattern generator located in the brainstem and more rostral spinal cord and constructed from inhibitory and excitatory premotor interneurons and motoneurons (Roberts and Perrins 1996; Roberts et al. 1997). When ®ctive swimming is initiated in immolised tadpoles it can last from a few seconds to one or more minutes. Intracellular recordings from either spinal premotor interneurons or motoneurons show that each of these neuron types ®res a single action potential on each cycle of swimming. Following initiation the swimming frequency is usually high but then gradually slows as swimming continues (Fig. 1B; Kahn and Roberts 1982). Some of this slow-down in swimming may be due to the extracellular accumulation of adenosine (Dale and Gilday 1996). However, frequency can also ¯uctuate up or down spontaneously, and be actively increased by stimulation of the skin (Sillar and Roberts 1988, 1992, 1993) or the pineal eye (Jamieson 1997). Could changes
754
in the level of synaptic excitation within the central pattern generator be responsible for such changes in frequency? Recent studies on the hatchling Xenopus tadpole have used the microperfusion of antagonists onto the recorded neuron to analyse synaptic excitation during ®ctive swimming (Perrins and Roberts 1995a,b,c). They suggest that tadpole motoneurons receive three kinds of excitatory input: glutamatergic from premotor interneurons, nicotinic cholinergic and electrotonic from neighbouring motoneurons. How could these excitatory inputs diminish as swimming continues and be changed by external stimulation during swimming? Since the neurons in the central pattern generator all ®re a single impulse on each cycle at this stage of development, there is no evidence that excitation can change as a result of presynaptic interneurons ®ring variable numbers of impulses on each cycle. The other main possibility is that the number of excitatory presynaptic neurons ®ring on each cycle changes. If more ®re, more transmitter will be released which will increase the frequency of the central pattern generator. Is there any evidence for changes in the ®ring of spinal neurons as frequency changes during swimming? Sillar and Roberts (1993) have shown that motoneurons ®re action potentials very reliably throughout episodes of swimming over the whole range of frequencies. On the other hand, premotor interneurons ®re less reliably as frequency falls (see also Perrins and Roberts 1995b). We therefore proposed that swimming frequency was partly controlled by the number of premotor excitatory interneurons active on each cycle which would determine the strength of the excitatory synaptic drive (Sillar and Roberts 1993). In this proposal spike failure in an increasing number of interneurons accounts for some of the fall in frequency as an episode progresses and the recruitment of more interneurons leads to the transient increase of swimming frequency following skin stimulation (Sillar and Roberts 1993; Wolf and Roberts 1995). It was only later that the cholinergic component of central pattern generator excitation was discovered and it became clear that global blocking of this cholinergic component could reduce the frequency of swimming (Perrins and Roberts 1995b). If cholinergic antagonists can change swimming frequency, do motoneurons normally play a part in frequency control or is this primarily the role of the premotor glutamatergic interneurons? Our aim was to resolve these questions by testing two predictions: (1) Since most motoneurons ®re action potentials very reliably at all frequencies, the population of motoneurons ®ring on each cycle should not change. As a consequence, the cholinergic excitation and electrical coupling components coming from motoneurons will not change with swimming frequency and will therefore not contribute directly to changes in frequency. (2) The glutamatergic component of excitation during swimming comes from premotor interneurons whose ®ring probability is thought to decrease with frequency. The
glutamate component should therefore decrease and lead to a fall in swimming frequency (Sillar and Roberts 1993). To test these predictions we recorded from motoneurons during ®ctive swimming episodes and microperfused speci®c antagonists locally onto the recorded motoneuron to reveal the glutamatergic, cholinergic and electrotonic components. We measured these to see if their amplitudes changed with frequency during swimming. Some parts of this work have been published in preliminary form (Zhao and Roberts 1996).
Materials and methods Xenopus tadpoles of developmental stage 37/38 (Nieuwkoop and Faber 1956) were anaesthetised with 0.1% MS-222 (3-aminobenzoic acid ethyl ester, Sigma) and pinned to a Sylgard table in a bath constantly perfused with saline containing (mmol l)1): NaCl 115, KCl 3, CaCl2 4, NaHCO3 2.4, MgCl2 1, HEPES 10, pH 7.4 at 18±22 °C. Under a stereomicroscope, a slit along the dorsal ®n was made using etched tungsten pins and the animal was then immobilised with a-bungarotoxin (Sigma, 10 lmol l)1 in saline) for up to 20 min. After immobilisation the animal was moved back to the bath and repinned with its right side up (Fig. 1A). Some skin and muscles covering the right spinal cord were removed and a neuron was recorded in the ventral 25% of the spinal cord where previous morphological studies have shown that motoneurons are densely packed (Roberts and Clarke 1982; Soe and Roberts 1982). A ventral root recording on the right side was used in all experiments to con®rm that ®ctive swimming occurred when motoneurons were active. The activity of presumed motoneurons was recorded intracellulary with glass microelectrodes with resistances of 150±250 MW when ®lled with 3 mol l)1 potassium acetate. The signals were ampli®ed and sent to a PC via a CED (Cambridge Electonics Design) interface for storage and o-line analysis using Sigavg software (CED). The signal was also monitored on an oscilloscope and recorded with a Racal FM tape recorder for back-up. Fictive swimming was initiated by dimming the illumination and monitored by recording ventral root potentials through a hole made on the skin overlying the 2nd or 3rd post-otic segment. Synaptic antagonists were applied as described previously (Perrins and Roberts 1995a). In outline, a multi-barrelled, switchable microperfusion system with a ®ne tip opening was used to direct a 50- to 80-lm-diameter stream of saline at the 70- to 100-lmdiameter spinal cord. Once a stable recording was made, this stream was centred on the recorded motoneuron and positioned about 30 lm from the cord to bathe the region around the motoneuron continuously. Excitatory antagonists were applied by switching the supply of perfusate. Since motoneurons have very limited longitudinal dendrites (less than 30 lm spread, Roberts and Clarke 1982), we are con®dent that the drugs applied will reach all synapses onto the dendrites and soma of the recorded motoneuron. The exact area aected by drug application cannot be assessed accurately for each recording but should be similar. In this way synapses onto the motoneuron were blocked but the operation of the spinal central pattern generator rostral to the application site was not aected. Strychnine (Stry 2 lmol l)1, Sigma) was always present in the microperfusion to block the glycinergic mid-cycle inhibitory post-synaptic potentials (IPSPs) and make measurement of the excitatory components possible. During swimming we measured the instantaneous frequency and the amplitudes of the tonic depolarisation and the phasic excitatory post-synaptic potentials (EPSPs) underlying spikes as previously described (Perrins and Roberts 1995a; see also Fig. 1C). All values quoted are means SEM. Regression analysis was performed using Microsoft Excel and P < 0.05 was the criterion for signi®cance unless otherwise stated.
755
Results De®ning the parameters to be measured during ®ctive swimming When an electrical pulse is given to the tadpole skin, or the illumination is dimmed, ®ctive swimming can be initiated (Fig. 1B). During ®ctive swimming motoneurons ®re a single action potential on each cycle. If no further stimuli are given, the frequency is highest at the beginning (above 20 Hz) but becomes lower as swimming continues, and ®nally, when it reaches about 10±12 Hz, swimming stops. If glycinergic mid-cycle inhibition is blocked by microperfusion of 2 lmol l)1 strychnine over the recorded neuron, then the synaptic excitation to motoneurons can be measured (Fig. 1C). It can be separated into a steady tonic depolarization (TD) which is measured as the lowest point between spikes and a phasic, on-cycle EPSP which underlies each spike and is measured as an in¯ection on the rising phase of the spike. The tonic depolarisation is thought to arise by the cycle-to-cycle summation of the slow N-methyl-D-aspartic acid (NMDA) components of glutamate-mediated excitation (Dale and Roberts 1985). Both components of excitation decrease as frequency decreases (Fig. 1C) but the change in tonic depolarisation is more obvious. Measuring the components of excitation as a function of swimming frequency Fig. 1A±C Fictive swimming in a motoneuron and the measurement of excitatory synaptic input. A Diagram of the stage 37/38 tadpole with skin and muscle removed on the right side to expose part of the spinal cord and allow intracellular recording from a motoneuron (mn). Synaptic antagonists were applied locally via a ®ne microperfusion jet positioned very close to the microelectrode and were carried away by the constant saline ¯ow through the bath. Ventral root potentials (VRP) were recorded through a small hole in the skin with a glass suction electrode. B An episode of ®ctive swimming initiated by dimming the illumination recorded from a motoneuron located at 7th post-otic segment. The resting potential was ±74 mV. The mid-cycle inhibition was blocked by microperfusion of 2 lmol l)1 strychnine. Swimming starts with higher frequency but as it continues the frequency gradually falls. The graph shows the change in instantaneous frequency over the whole episode. C Periods of swimming (a, b, c) sampled from the episode in A at an expanded time scale where the tonic depolarization and on-cycle EPSPs, measured as indicated in c, are higher at the beginning, and decrease as swimming continues. Lower dotted line shows resting potential before swimming started Drugs used were: strychinine (glycine antagonist, 2 lmol l)1, Sigma, USA); dihyro-b-erythroidine (DHbE), nicotinic ACh antagonist, 10 lmol l)1, Research Biochemical International, USA); kynurenic acid (non-selective glutamate anatgonist, 2 mmol l)1, Sigma, USA); 6-nitro-7-sulphamoylbenzo[f]-quinoxaline-2,3-dione (NBQX, selective (RS)-a-amino-3-hydroxy-5-methyl-4-isoxazolepropionic acid (AMPA) antagonist, 5 lmol l)1, Tocris Cookson, UK); cadmium chloride to block all chemical synaptic transmission (100 lmol l)1, Sigma, USA). Kynurenic acid was ®rst dissolved in 1 Eq NaOH at 100 mmol l)1 and NBQX in dimethyl sulfoxide (DMSO) to 20 lmol l)1. The other drugs were dissolved in water for stock solutions and were then diluted with saline to their ®nal concentrations before application.
Since Tunstall and Roberts (1994) have shown that excitatory synaptic drive decreases caudally, we selected motoneurons in more rostral positions between the 5th and 7th segments where the EPSP and tonic depolarisation would be large. In almost all swimming episodes that we recorded, the frequency and the motoneuron synaptic excitation changed more rapidly at the beginning than in the middle or near the end (Figs. 1B, 2A). To reveal the relationships between frequency and synaptic excitation, we therefore selected the ®rst 600± 1000 ms of ®ctive swimming for analysis in seven motoneurons. In the presence of 2 lmol l)1 strychnine, motoneurons ®re even more reliably than in the normal saline (Perrins and Soe 1996). An example is shown in Fig. 2A. Following stimulation (dimming the illumination) excitation rises quickly from the resting potential and then gradually drops as the swimming progresses. Action potentials are ®red on each cycle and the frequency is higher earlier than later in the episode. When the amplitudes of tonic depolarisation and EPSP were plotted, they both decreased with swimming frequency and this decrease was shown to be signi®cant using regression analysis (Fig. 2A). The same analysis showed a clear correlation between frequency and the amplitudes of tonic depolarisation and EPSP in all seven neurons. The dierent components of excitation were then examined. When DHbE was microperfused over the
756
motoneurons to block the nicotinic acetylcholine (nACh) component, they continued to ®re action potentials and both components of excitation decreased signi®cantly with frequency (Fig. 2B). In DHbE tonic depolarisation changed with frequency in all seven neurons and the EPSP changed in six out of seven neurons. During the application of kynurenate +NBQX to block the glutamate component, all six neurons recorded stopped ®ring spikes, but the underlying rhythmic activity was still clear (Fig. 2C). The amplitudes of both tonic depolarisation and EPSPs did not change signi®cantly with frequency (Fig. 2C). When all chemical synaptic input was blocked by microperfusion of Cd2+ there was similarly no change in the amplitudes of tonic depolarisation and EPSPs with frequency (Fig. 2D) in six out of seven neurons. One of the seven neurons showed EPSPs whose amplitude changed with frequency. At the very beginning of a swimming episode, the sensory input which initiates swimming may contribute to the amplitudes of tonic depolarisation and EPSPs. To exclude this possibility, we therefore measured tonic depolarisation and EPSPs between 1 s and 1.5 s after the beginning of the episode in 11 neurons. The results from these measurements are shown in Fig. 3. The regression lines are shown solid where the slope was signi®cant and dotted where it was not signi®cant. This second dataset measured later in the swimming episode shows similar trends to those illustrated by the examples in Fig. 2. Provided that the glutamate component is present, as in control and 10 lmol l)1 DHbE, both EPSPs and tonic depolarisation change with frequency. However, when the glutamate component is blocked, by kynurenate and NBQX (n 6) or by Cd2+, then neither potential change signi®cantly with frequency.
Discussion Fig. 2A±D In¯uence of frequency on components of synaptic excitation at the start of swimming. A±D In each part the left side is a record from the motoneuron at the start of the swimming episode evoked by dimming the light with resting potential marked by dotted line, and the right side is a graph of the amplitudes of the tonic depolarisation (TD) and excitatory post-synaptic potential (EPSP) as a function of swimming frequency. A In 2 lmol l)1 strychnine, the tonic depolarisation and EPSPs decrease with the frequency (n 14, regression analysis, r 0.921 and 0.798, P < 0.001 for both tonic depolarisation and EPSPs). B In 10 lmol l)1 dihyro-b-erythroidine (DHbE) to block the nicotinic acetylcholine component, the tonic depolarisation and EPSPs are slightly reduced but both still decrease with frequency (n 21, r 0.806 and 0.731, P < 0.001 for both tonic depolarisation and EPSP). C In 2 mmol l)1 kynurenate (Kyn) and 5 lmol l)1 6-nitro-7-sulfamoylbenzo(f)-quinoxaline-2,3-dione (NBQX) to block glutamate receptors (GlutRs), the tonic depolarisation and EPSPs are dramatically reduced, action potentials are no longer present, and the remaining tonic depolarisation and EPSPs are not frequency dependent (n 16, r 0.471 and 0.120, P > 0.05 for both tonic depolarisation and EPSPs). D In Cd2+ to block all chemical synaptic transmission, the remaining electrotonic component of tonic depolarisation and EPSPs did not change with frequency (n 22, r 0.144 and 0.261, P > 0.05). A, B, and D were from a single neuron but C was from another neuron
Our results show that blocking nACh receptors with the speci®c antagonist DHbE did not prevent the excitatory input to motoneurons changing with swimming frequency. On the other hand, if glutamate receptors (GlutRs) are blocked with the antagonists kynurenate plus NBQX, then the remaining excitatory synaptic excitation does not change with frequency. These observations con®rm our prediction that the nicotinic acetylcholinergic component of excitation that comes from motoneurons does not contribute directly to the normal slow-down of the locomotor central pattern generator as swimming progresses. The results strongly imply that the slow-down of swimming is partly due to a decrease in glutamatergic synaptic excitation from premotor interneurons. Another possibility is that the fall in frequency and reduction in the glutamate component is due to desensitisation which is known to occur in AMPA receptors (Jones and Westbrook 1996; Lukasiewicz et al. 1995). However, desensitization is unlikely to be the whole explanation as sensory stimulation
757 Fig. 3A±D Graphs of the amplitudes of excitation (tonic depolarisation and EPSPs) as a function of swimming frequency in control (with strychnine 2 lmol l)1) and in the presence of excitatory antagonists. The amplitudes of tonic depolarisation and EPSPs were measured 1±1.5 s after swimming began for about 1 s. Solid lines represent regression lines with a statistically signi®cant slope and dotted lines those without. A In control with strychnine, the tonic depolarisation and EPSPs changed with frequency in 10 out of 11 neurons. B In DHbE to block nAChR, most of the tonic depolarisation (10 out of 11 neurons) and EPSPs (8 out of 11 neurons) still changed with the frequency. C In kynurenate + NBQX (6 neurons) where the glutamate components were blocked, the amplitudes of tonic depolarisation and EPSP did not change with frequency. D In Cd2+, only one neuron showed a correlation between tonic depolarisation and frequency and EPSPs did not change with the frequency
during swimming can produce a sudden rise in frequency (Sillar and Roberts 1988, 1992). Our results leave us with a question to answer: what are the functions of the nACh and electrical components of excitation during swimming? Perrins and Roberts (1995b) have shown that nicotinic antagonists reduce the ®ring reliability of premotor interneurons during swimming, and when applied to the whole spinal cord can
reduce swimming frequency. They concluded from this evidence that part of the excitation to the spinal central pattern generator is nACh mediated and almost certainly comes from motoneurons. The present results suggest that this cholinergic component of excitation does not vary with frequency. Since motoneurons show the same activity at all frequencies, the excitation that they produce will remain stable and not change. We
758
propose that the cholinergic and electrotonic components of excitation ensure that whenever swimming occurs there is a basic stable level of central excitation. A variable amount of glutamate-mediated excitation sums with this base level and it is this variable glutamatemediated component which controls frequency (Sillar and Roberts 1993). These synaptic mechanisms contribute to the natural fall in frequency as swimming proceeds. They are important as they can lead to increases in frequency following stimulation during swimming, and are superimposed onto other non-synaptic mechanisms which also contribute to the slow down and termination of swimming (Dale and Gilday 1996; Dale and Kuenzi 1997). Acknowledgements This work was supported by the Wellcome Trust. The authors wish to thank Drs S. R. Soe and E. Wolf for their advice, Drs S. R. Soe and Ray Perrins for commenting on the manuscript, and Derek Dunn and Linda Teagle for their excellent technical assistance. The experiments described in this paper comply with the laws of the United Kingdom.
References Dale N, Gilday D (1996) Regulation of rhythmic movements by purinergic neurotransmitters in frog embryos. Nature (Lond) 383: 259±263 Dale N, Kuenzi FM (1997) Ion channels and the control of swimming in the Xenopus embryo. Prog Neurobiol 53: 729±756 Dale N, Roberts A (1985) Dual-component amino-acid-mediated synaptic potentials: excitatory drive for swimming in Xenopus embryos. J Physiol (Lond) 363: 35±59 Jamieson D (1997) Synaptic transmission in the pineal eye of young Xenopus leavis tadpoles: a role for NMDA and non-NMDA glutamate and non-glutaminergic receptors? J Comp Physiol A 181: 177±186 Jean A (1984) Control of the central swallowing program by inputs from peripheral receptors: a review. J Auton Nerv Syst 10: 225±233 Jones MV, Westbrook GL (1996) The impact of receptor desensitization on fast synaptic transmission. TINS 19: 96±101 Kahn JA, Roberts A (1982) The central nervous origin of the swimming motor pattern in embryos of Xenopus laevis. J Exp Biol 99: 185±196 Lukasiewicz PD, Lawrence JE, Valentino TL (1995) Desensitizing glutamate receptors shape excitatory synaptic inputs to tiger salamander retinal ganglion cells. J Neurosci 15: 6189± 6199 McClellan A, Grillner S (1983) Initiation and sensory gating of ``®ctive'' swimming and withdrawal responses in an in vitro preparation fo the lamprey spinal cord. Brain Res 269: 237±250
Nieuwkoop PD, Faber J (1956) Normal tables of Xenopus laevis (Daudain). North Holland, Amsterdam Perrins R, Roberts A (1995a) Cholinergic and electrical motoneuron-to-motoneuron synapses contribute to on-cycle excitation during swimming in Xenopus embryos. J Neurophysiol 73: 1005±1012 Perrins R, Roberts A (1995b) Cholinergic contribution to excitation in a spinal locomotor central pattern generator in Xenopus embryos. J Neurophysiol 73: 1013±1019 Perrins R, Roberts A (1995c) Cholinergic and electrical synapses between synergistic spinal motoneurons in the Xenopus laevis embryo. J Physiol (Lond) 485: 135±144 Perrins R, Soe SR (1996) Local eects of glycinergic inhibition in the spinal cord motor systems for swimming in amphibian embryos. J Neurophysiol 76: 1025±1035 Roberts A, Clarke JDW (1982) The neuroanatomy of an amphibian embryo spinal cord. Philos Trans R Soc Lond B 296: 195±212 Roberts A, Perrins R (1996) Positive feedback as a general mechanism for sustaining rhythmic and non-rhythmic activity. J Physiol (Paris) 89: 241±248 Roberts A, Soe SR, Perrins R (1997) Spinal networks controlling swimming in hatchling Xenopus tadpoles. In: Stein PSG, Grillner S, Selverston AI, Stuart DG (eds) Neurons, networks and motor behavior. MIT Press, Boston, pp 83±89 Robertson GA, Mortin LI, Keifer J, Stein PSG (1985) Three forms of the scratch re¯ex in the spinal turtle, central generation of motor patterns. J Neurophysiol 53: 1517±1534 Sillar KT, Roberts A (1988) A neuronal mechanism for re¯ex modulation during locomotion. Nature (Lond) 331: 262±265 Sillar KT, Roberts A (1992) Phase-dependent modulation of a cutaneous sensory pathway by glycinergic inhibition from the locomotor rhythm generator in Xenopus embryos. Eur J Neurosci 4: 1022±1034 Sillar KT, Roberts A (1993) Control of frequency during swimming in Xenopus embryos: a study on interneuronal recruitment in a spinal rhythm generator. J Physiol (Lond) 472: 557±572 Soe SR, Roberts A (1982) Tonic and phasic synaptic input to spinal cord motoneurons during ®ctive locomotion in frog embryos. J Neurophysiol 48: 1279±1288 Stein PSG, Grillner S, Selverston AI, Stuart DG (1997) Neurons, networks and motor behavior. MIT Press, Boston Tunstall MJ, Roberts A (1994) A longitudinal gradient of synaptic drive in the spinal cord of Xenopus embryos and its role in co-ordination of swimming. J Physiol (Lond) 474: 393±405 Wolf EW, Roberts A (1995) The in¯uence of premotor interneurone populations on the frequency of the spinal pattern generator for swimming in Xenopus embryos: a simulation study. Eur J Neurosci 4: 671±678 Zhao F-Y, Roberts A (1996) Antagonists unmask excitatory synaptic drives to motorneurons during swimming in Xenopus tadpoles. Abstracts, Society for Neuroscience 26th Annual Meeting, Washington DC, p 1375
