Lf)

000

,C E

0 -00000 L,

* 0.

0J* 0

0S .U 0 ow.

0 00 . 00* S 0

0 *g 0)0'0 0

*0 Z*. .

0 000 'U

0~ 0s-.

C.2T OLU h

0~ **. h

r. -

z %a

00 0 ..=

00. IA

-~ 0SO. 0 .I

55@

A. 0.000:.. 0

-j

Re:,ro: iced by theCLEARING HO0US5E

to, evdvrjl Scwnfifiz & TeOM7a 1

Informatior Spn-ngfield Va. 22151

Report No. IITRI-L6021-20

DEVELOPMENT OF AN ORALLY EFFECTIVE INSECT REPELLENT

Annual Progress Report

Philip Kashin

December 1969

U. S. Army Medical Research and Development CommandOffice of The Surgeon General

Washington, D. C. 20315

Contract No. DA-49-193-MD-2281

IIT RESEARCH INSTITUTE10 West 35th Street

Chicago, Illinois 60616

ABSTRACT

Studies were continued in efforts to elucidate the basicphysiological mechanisms that may be involved in the attrac-tion of mosquitoes to their warm-blooded hosts. Experimentalobservations were made to test the -validity of the gamma-amiEobutyric acid (GABA)-carbon dioxide (C02 ) hypothesisdescribed in previous reports. This hypothesis could ex-plain the neurochemical interactions that govern a mosquito'sbehavior in the presence of its host. It was postulated thatthe formation of carbamino-GABA underlies the activating ef-fects of CO2 on mosquitoes.

The isolated central nervous system of the cockroach,Periplaneta americana, was utilized as a model system totest the effects of carbamino-GABA and other compounds onthe nervous system of an insect. GABA inhibited the spon-taneous firing rate of this preparation in the absence ofC02, but stimulated it in its presence. Gamma-hydroxybutyricacid, a substance which cannot form a carbamino compound,was inhibitory both in the absence and presence of CO Achemical analogue of carbamino-GABA, N-acetyl GABA G),stimulated the preparation. Evidence was obtained showingthat -NAG and GABA probably compete for the same site on thesynaptic membrane. These results have important implicationsin terms of the mechanism by which C02 stimulates specializedinsect nervous tissue. Further studies are in progress.

The evaluation of selected compounds for mosquito repel-Iency by the electronic recording method was continued. Com-pounds were selected on the basis of our past approach, i.e.that volatile chemical analogues of GABA, in terms of elec-tronic distribution, would affect the mosquito's responsesto its host.

A statistically designed series of tests were performedto test the validity of some basic assumptions that are in-volved in using the electronic method for the assay of ra-pellents. The results of these tests largely substantiatedour confidence in this method.

FOREWORD

This is Report No. IITRI-L6021-20 (Annual ProgressReport) on IITRI Project L6021, entitled "Developmentof an Orally Effective Insect Repellent." The reportcovers the period frcm November 1, 1968 through October 31,1969.

This project is being sponsored by the U. S. ArmyMedical Research and Development Command, Office of TheSurgeon General, Washington, D. C. 20315, under ContractNo. DA-49-193-MD-2281 and is being conducted by lITResearch Institute, Technology Center, Chicago, Illinois60616. Previous work under this contract was conductedby IT Research Institute from May 1, 1962 throughOctober 31, 1968.

The project leader for this program is Mr. PhilipYashin, under the administrative supervision of Dr. MaryHenry. The statistical analyses were performed by Mr.Merl L. Kardatzke, who also dev.-sed the computer programfor determining the repellency index and the statisticalconfidence limits for the test compounds. Valuable sug-gestions and discussions for the physiological phases ofthe work were contributed by Dr. William F. Danforth,Bio'.ogy Department, Illinois Institute of Technology.Dr. James C. Lambert of the Electronics Division of IITR search Institute suggested the modified recording ar--angement for the electrophysiological recordings des-eribed below. Mr. Anthony M. Gross provided technicalassistance in theassay of repellents.

In conducting the research described in this report,the investigator adhered to the "Guide for LaboratoryAnimal Facilities and Care" as promulgated by the Committeeon the Guide for Laboratory Animal Resources, NationalAcademy of Sciences - National Research Council.

The citation of any trade names in this report doesnot constitute an official endorsement or approval of theuse of such commercial hardware or software.

All repellency test data are recorded in IITRI LogbooksC18467 and C19669. The computer output sheets also formpart of our permanent records.

3

TABLE OF CONTENTSPage

I. Introduction 9

II. Physiological Studies 10

A. Modifications of Electrophysiological 10Methods

B. Modified Physiological Saline Solution 13C. Experimental Results and Discussion 14

III. Statistical Tests of "Bitometer" Method 28

IV. Repellency Assay of Selected Compounds 39

V. Summary 45

A. Electrophysiological Investigations 45B. Studies of Repellency Assay by the 46

Electronic MethodC. Tests of Repellents 46

VI. Future Investigations 46

A. Further Investigations of the GABA Hypothesis 46B. Search for New Repellents 47C. Maintenance of Mosquito and Cockroach 47

Colonies

VII. Conclusions 47

Literature Cited 49

DD Form 1473

4

LIST OF TABLES

Table Page

I Randomized Block Design for Testing Effects 31of Mosquito Age, Time of Day, and OtherParameters on the Mosquito Biting Rate onRepellent-Treated and Control Cases

Repellency Indices with Statistical Confidence 32Limits for Age and Time of Day Effects inControl and Repellent-Treated Cases

3 Analysis of Variance of Repellency Index for 36Controls

4 Analysis of Variance of Repellency Index for 36Repellent-Treated Cases

5 Analysis of Variance for Engorgement (E) of 37Controls

6 Analysis of Variance for Percent Biting Time 37(P) of Controls

5

LIST OF FIGURES

Figure Pg

I Revised Method for Recording Spontaneous 12Bioelectric Activity of the IsolatedCentral Nervous System of the Cockroach

2 Effects of (a) CO2 , (b) GABA 10-6 M (c) 15GABA i0-6 M + 10/ C02 , (d) GABA 10 - 4 m,and (e) GABA 10- 4 M + 10% C02 on theSpontaneous Discharge Rate of the IsolatedCockroach CNS

3 Effects of (a) GABA 5 x 10-4 M, (bj GABA 165 x 10 - 4 M + 10% CO2, (c) GABA i M,(d) GABA i0-3 M + 10 7 C02, (e) GABA 5 x 10-3 Mand (f) GABA 5 x 10-1 M + 10% C02 on theSpontaneous Discharge Rate of the IsolatedCockroach CNS

4 Effects of (a) GABA 10-2 M + 10% C02 and 17(b) Saline + 10% CO2 (Before and After theAddition of GABA 5 x 10-3 M), on theSpontaneous Discharge Rate of the IsolatedCockroach CNS

5 Effects of (a) GHB 10 - 6 M, (b) GH3 I0-6 M 20+ 105% C02 , (c) GHB 10- M, (d) GHB 10- 4 M+ 10/ C0 2 , (e) GHB 10 "3 M, and (f) GHB 10 - 3 M+ 10%, CO2 on the Spontaneous Discharge Rateof the Isolated Cockroach CNS

6 Effects of GHB 5 x 10 - 3 M 2 (b) GHB 5 x 10- 3 M 21+ 10% CQ2 , (c) GHB 5 x 10 M, (d) GHB5 x 10- M + 10% CO2, and (e) Saline + 10% CO2after GHB Treatment on the Spontaneous DischargeRate of the Isolated Cockroach CNS

7 Effects of Various Concentrations of NAG on 23the Spontaneous Discharge Rate of theIsolated Cockroach CNS

6

LIST OF FIGURES (cont.)

Figure pag

8 Effects of Adding (a) NAG 10 - 4 M, (b) NAG 10 - 3 M, 24(c) NAG 5 x 10 - M, and (d) NAG i0 - 2 M to theCocroach Preparation During Treatment with10 -4 M GABA

9 Effects of Adding (a) NAG 10 - 3 M, (b) NAG 255 x 10 - 3 M, and (c) NAG 10-2 M to the CockroachPreparation during Treatment with 10-3 M GABA

10 Effect of Mosquito Age on the Repellency Index 33of Control and Low-Level Repellent-Treated Cases

11 Effect of Time on Day on the Repellency Index 34of Control and Low-Level Repellent-Treated Cases

7

DEVELOPMENT OF AN ORALLY EFFECTIVE INSECT REPELLENT

I. INTRODUCTION

The objectives of this program are to develop bettertopical insect repellents than those currently available,and if possible, to develop an insect repellent that iseffective when administered orally. Such repellents shouldafford more uniform and longer-lasting protection from bitesof hematophagous insects. Success in this undertaking couldresult in n significant reduction of htuman suffering fromdisease and discomfort caused by insect bites.

In order to accomplish these goals, studies were con-tinued during this year to elucidate the basic physiologicalmechanisms that may be involved in the attraction of hemato-phagous insects to their warm-blooded hosts. An understandingof this mechanism would permit the premeditated design ofchemical compounds that could confuse and interfere with theresponses of a hematophagous insect to its host.

ihe experimental ratienale and approaches followed inthe work during this report period were largely establishedin previous years (ref. 1 to 5). Ganrna-aminobutyric acid(GABA), a substance known to cause inhibition in the trans-mission of nerv.ous impulses across certain synaptic struc-tiires, was found in mosquitoes (ref. 2 to 5). It was hypo-thesized that the interactions of GABA with carbon dioxide,heat, and t,ater vapor form the basis of mosquitoes' attractionto their hosts. Further evidence supporting the hypothesisthat the formation of carbamino-GABA underlies the activatingeffects of CO on mosquitoes was obtained from the effects ofCO-, GABA, an3 gamina-hydroxybutyric acid, and N-acetyl-GABA(a-chemical analogue of GABA-C02) on the spontaneous firingrate of the isolated central nervous system of the Americancockroach, Periplaneta americana (L).

The testing of volatile chemicals for repellency that areGABA analogues in terms of electronic configuration was con-tinued. Compounds were selected on the basis of our pastapproach, i.e., that chemical molecules that have an elec-tronic distribution similar to that of GABA or its analogueswould effect the mosquito's responses to its host.

9

In evaluating candidate compounds for mosquito repellency,a method is utilized wherein the bite of a mosquito is elec-tronically detected and recorded. The electronically recordeddata, together with the percentage of mosquitoes that engorgedblood during a test are compared with the same parameters de-rived from controls in a computerized statistical analysis toyield statistical confidence limits for the repellent efficacyof test compounds.

A statistically designed series of tests were performed totest the effects of age, time of day, and other parameters onthe biting rate of mosquitoes on control and repellent-treatedanimals using the electronic method. This was done in orderto test the assumption that the variables affecting the bitingrate on untreated controls will also affect the repellent-treated cases in a parallel way.

This annual report details further progress in our in-vestigations in these areas.

II. PHYSIOLOGICAL STUDIES

A. Modifications of Electrophsioloical ilethods

The previous method for recording the spontaneous elec-trical activity from the 3rd thoracic ganglia of the isolatedcentral nerve cord of the American cockroach, Periplanetaamericana (L.) (ref. 4) was modified to give better recordingsfrom this preparation. We previously used a single-ended system,where one electrode was the ground. This was not the optimalarrangement for collecting the data, since the system was notbalanced electrically. This resulted in an increased noiselevel and loss of sensitivity. Stray AC signals also provedto be a problem, even in the electrically shielded enclosurethat is used for this work since some signals arose from theAC equipment within the enclosure. The stray AC signals wereeliminated by partially shorting the recording electrode tothe ground through a 10,000 ohm resistor. This procedure causedan even further reduction in sensitivity.

10

The circuitry is now changed so that the reference electrodeis a true reference electrode with regard to the recording elec-trode, and the data-gathering electronic system is utilized ina true push-pull mode. A one megohm potentiometer was installedbet-een the ground, the reference electrode, and the recordingelectrode as shown in Figure 1. The electrical balance betweenthe reference and recording electrodes can be adjusted with thep.'tentiomerer in such a way that stray AC signals are virtuallyeliminatped, while the signal strength is enhanced.

This i-3 done as follows. The reference electrode is im-mersed into the saline solution in the 5 cc beaker, while therecording electrode with the ganglion chain (usually consistingof the 3rd thoracic ganglia and 3 or 4 abdominal ganglia) ishanging freely above the solution and not touching it. Thepotentiometer is turned until the AC signal is eliminated, asdetermined from the oscilloscope trace. At this point the elec-trodes are balanced and the electronic counter can be calibratedas described below.

it is i:nporzant to standardize che method of setting the sen-sitivity level of the electronic digital counter which talliesthe spontaneous action potentials from the nerve cord over the5-sec counting interval. If the level is set too low, data will.be missed; if set too high, too much noise will be recorded.The following method is used to set the sensitivity level of theountler. With the electrodes shorted in the saline solution,

'i[e counter sensitivity control is set to just barely show nocoupMs when the preamplifier amplification control is set onestep higher (more sensitive) tLhnn the setting that is useddu:ing the experiment. (Each amplification step on the Grass'Model P-4 preamplifier that is used for this work is a multiple.- 1.5). The preamplifier amplification control is then set to:be level at which the experimental data will be taken (i.e. one

Ptep lowe. than the setting at which the zero level was calibrated).Using this procedui-e, a signal to noise ratio of approximately1.5:1 is obtained, which is very satisfactory for this work.

Treatment of the nerve cord with chemicals has also beenstandardized. The test solution is placed in the 5 cc beaker,the beaker is raised by means of the jack so that the 3rdthoracic ganglia and recording electrode are completely emersedini the test solution for 5 sec. The beaker is then slowly lowered,until the surface tension just breaks from the recording electrodeholding the ganglia. When the surface tensions breaks, the elec-trodes are unshorted, and the data gathering can proceed. The

i1

H u

u0

00P

H 0

zZ -

Z 0 )0

0U H .3

F1 0 H

I0

w- 0Ix

00

0 Zc

4.

Iz

wo

I1

moment of unshorting of the electrodes (when the sarface ten-sion breaks) can be observed visually, but is usually deter-mined when the oscilloscope trace suddenly changes. Thismethod has proven to be reproducible and accurate. Therefore,the amount of the nerve ccrd which is removed from the solutionis carefully controlled. This is iportant, since the recordeddischarge rate increases as more of the nerve cord is removedfrom the solution.

When the preparation is treated with a test solution, the5 cc beaker containing the saline in which the preparation issuspended is lowered by means of the jack, and the salinebeaker is replaced with the beaker containing the test solu-tion. The test beaker is then raised to the preparation, andthe entire preparation is immersed in the test solution for5 sec. After this Lime, the 5 cc beaker is lowered until thesurface tension breaks from the recording electrode as describedabove, and data is recorded. In the following figures, an arrowpointing to a double circle designates the last point of thedata before the test solution was added. Therefore, immediatelyaftear this point there is actually a discontinuity during whichthe above-mentioned operations were performed.

B. Modified Physiological Saline Solutions

Another modification in the experimental procedure wasmade. Bicarbonate buffer was omitted from the saline solu-tion and a new buffer substituted. This buffer, N-2-hydroy-etlilpiperazine-N'.-2-ethane-sulfonic acid (HEPES), has a pKvalue of 7.55, and has been successfully used in tissue cultureand biochemical work. This buffer was used since bicarbonatebuffer can give rise to C02. The new buffer contributes nobicarbonate to the system, and has a pK value which is in thephysiologic range.

The buffered saline solution we now use has the following

composition. All quantities are in terms of grams/liter.

NaCl 9.1

KCI 0.9

CaC12 5.07

MgC12.6H20 0.8

Glucose 4.0

HEPES buffer 2.383 (0.01 M)

13

The completed solution is carefully titrated to pH 7.60.A concentrated stock solution (10x) is diluted and titratedas needed, The above saline is a modification of that des-cribed by Smit et al (ref. 6), This buffered saline solutionkeeps the preparation in good condition for many hours. Thefollowirg experiments were carried out with all the modifi-cations described above.

C. Exoerimental Results and Discussion

1. Studies of the Effects of Carbon Dioxide, GABA, and Carbamino-GABA on the Insect Central Nervous System

Figure 2a shows the stimulatory effects of the sal ne solu-tion alone treated with CO2 . Figure 2b shows that 10P M GABAhas practically no neuroinhibito~y action, but when treatedwith 10% CO (Figure 2c) appears to be somewhat more stimulatorythan the saline-treated CO2 solution (Figure 2a). At a GABAconcentration of 10-4 M there may be a slight indication ofpossible inhibition (Figure 2d) but in the presence of '10% C02(Figure 2e) a high degree of stimulation is shown that seemsto be greater than that seen in Figure 2c.

Treatment of the preparation with 5 x 10- 4 M GABA (Figure3a) again possibly causes a slight immediate inhibition, butthe preparation is greatly stimulated when the solution istreated with 107 C02o

Application of 10 - 3 M GABA causes an immediate and pre-cipitous decline in the electrical activity of the preparation(Figure 3c) but this GABA solution becomes a potent stimulator

' hen treated with i07O C02 (Figure 3d). Continuing these ob-servations, it can be seen that treatment of the ganglia with5 x 1.0-3 M GABA again causes an immediate decline in the firingrate (Figure 3e) but when the solution is treated with, 10%/ CO2the control level of firing is maintained for almost a fullminute before inhibition sets _ (Figure 3f). Figure 4a showsthe effects of treating at 10 M GABA with 10' CO . Althoughthere wa5 an imediate decline in discharge rate when treatedwith 10- z M GABA (not shown), there is no decline for at least

30 sec after application of the C02-treated solution. All ofthe above experiments were performed on a single preparation.

14

•.' , g q

Im •°

(a)

oo

SO

(b) (c)

0%~ cot

"%M '-IDI

(e)

Figure 2

EFFECTS OF (a) CO 2 (b) GADA 10 6 M, (c) GABA 10-6 m + 10% co2

-4 -4-2

(d) GABA 10~ L, AND (e) CARA 10~ M + 107. C02 , ON THE SPONTANEOUS DISCHARGE RATEOF THE ISOLATED COCKROACH CNS

15

(a) (b)

Am.

.... "

E E O (a) b

5 x 1C

(c) (d)

~16

(e) (f)

Figure 3

EFFECTS OF (a) GAlA 5 x IO " H, (b) CAMA 5 x 10 -4 N + I0"/, CO2 ,

(c) GAlA 10 "3 H, (d) CAMA 10 " + 10O, CO2 , (e) GAlA 5 x 10 H ,

AND(f)G~l 5 K 1O~ C 2 ,ON TESPONITANEOUS DISCHARGE RATE

OF THE ISOLATED COCKROACH CS

16

700.

S/ p + /\ .10% c0

" A '\/o'L-0"

.4

~MIN

\

uAc

EFFECT OF.' (a AA1- 0,C6 N b I NE+1.C

1.) GAB (.. )

(BEOREAND ARTEADINO AA 5x 10-3 m*

00 ,17

;,ooI

(b)

Figure 4EFFECTS OF (a) GABA i0 -2 M +. 10L O 2 AND (b) SALINE +~ 10%, CO2

(BEFORE AND AFTER THE ADDITION OF GABA 5 x 10- M),

ON THE SPONTANEOUS DISCHARGE RATE OF THE ISOLATED COCKROACH CNS

17

It was also of interest to observe what the effects wouldbe on a preparation that was first treated with saline thathad CO2 bubbled into it, and into which an untreated GABAsolution was introduced to give a final concentration of5 x 10-3 M. It can be seen in Figure--4b that the preparationis stimulated after about 30 sec to a greater level than thatof C02 alone, followed by a gradual decline.

The amount of decline after the initial stimulation in yheC02-treated GABA solutions, having concentration of 5 x 10- MGABA or less, is to the control level, while the decline goesbelow control levels in the more concentrated C02 -treated GABAsolutions. This corresponds with the observations that inhi-bition at GABA concentration of 5 x 10-4 or below is hardlydiscernible, while inhibition at GABA levels of 10-3 M andhigher is clearly evident. Though a definite stimulation isseen immediately upon lifting thL ganglia from the solution,it appears that with time the full inhibitory action of GABAis established. This is probably due to the rapid escape ofC02 from the preparation after removal from the solution torecord the data. As the GABA concentration increases, the in-hibitory effect of GABA is established at the higher concentra-tions with the escape of C02 .

It should be mentioned that even if it were feasible toperform these experiments without removing the ganglia fromsolution and exposing it to air (i.e. if the problem of theshort-circuiting of the electrodes in solution could be over-come), it would still not be possible to do the experimentsin this way. This is because if the ganglia is continuouslyLmmersed in solution, it soon becomes anoxic and the spontaneousdischarges cease altogether.

2. Studies of the Effect of Carbon Dioxide and GHB on theInsect Central Nervous System

A parallel series of experiments were performed with gamma-hydroxy butyric acid (GHB). This compound was of interest be-cause it is a neuroinhibitor (ref. 7,8) although it is lessinhibitory than GABA (ref. 7,9). GHB has a chemical structuresimilar to that of GABA, with the exception that it possessesa gamma-hydroxy group instead of a gamma-amino group. SinceGHB does not possess an amino group, it cannot form a carbaminocompound with C02 and therefore can be used to test the hypo-thesis that carbamino-GABA formation is the mechanism by whichC02 stimulates the nervous system of the insect. If the mecha-nism of stimulation by CO is not that of carbamino formation,but instead through an ingependent receptor site for CO, thenthe results of the GHB experiment would be similar to t~ose ofthe GABA experiment.

18

Representative results of this experiment are shown inFigi,.es 5 (a-f) and 6 (a-e). It can be seen that the stimu-latory affects of CO2 steadily decline as the concentrationof GHB increases.

No inhibition is evident at GHB oncentrations of 10 - 6 M(Figure 5a), 10-4 M (Figure 5c), 10' M (Figure 5e) or5 x 10-3 M (Figure 6a). The stimulatory effects of C02 are ata maximunm at GHB 10-6 M + 10/ C02 (Figure 5b), decreases at aGHB concentration of 10 - 4 M + 10% CO2 (Figure jd), and are com-pletely abolished at GH1 c. ncentrations of 10" M + 10% CO2(Figure 5f) and 5 x 10"- M GHB + I10% CO2 (Figure 6b).

It is §een in Figure 6c that not until a GHB concentrationof 5 x 10-l M is reached can the inhibitory effects of GHB bediscerned, and that the addition cf 10% CO2 to this solution(Figure 6d) has little or no modifying effect on the inhibitionof this concentration of GHB. After this series of experimentsthe preparation was washed with saline, and then treated witha saline solution containing 10% CO2 (Figure 6e). It can bes--:. that :re preparation has completely lost its sensitivityto CO2 after the GHB treatment. Although not shown in thisseries of experiments, we previously reported (ref. 9) thatthe loss of sensitivity of the preparation to CO2 could bec . pletely reversed by subsequently treating the preparationwith GABA.

These findings lend strong support to the suggestion thatcarhamino-GABA formation is the mechanism by which C02 stimu-letes insect nervr.us tissue which is sensitive to the action.r CO 2 , apd to the hypothesis that carbamino-GABA is a neuro-stimu ator in certain specialized insect neuro-receptors.

It is also of interest to note that although GHB is about1 t:,- 2 orders of magnitude less inhibitory than GABA, and doesnot demonstrate any significant neuroinhibitory action at con-centrations less than about 5 x 10-2 M, the stimulatory effectsof CO2 show considerable decline in the presence of 10-4 M GHB(Figure 5d) and are totally abolished at a GHB concentrationof 10-3 M (Figure 5f). It thus appears that GHB even at re-latively low concentrations can successfully compete withendogenous GABA for the synaptic receptor site, and preventthe GABA-C02 formed from endogenous GABA (in the case of GHBsolutions treated with C02 ) from reaching the receptor site.This brings up the question of relative affinities for thesynaptic receptor site, and related to this, whether thereis one receptor site or more than one such site for Co, GABA,GABA-C02, and GHB. The following experiments represen Eourfirst attempts to answer these questiorts.

19

(a) (b)

/

" " '"\ , .f .----, .-"

,, 10"4 M

(c) (d)

(d GH 0 4 10 MU + 0 0,(c 0 M.",, 10"... ,O 2. (e) ,,B 10 3",A D f H 1 ",, 1%ON THE SPOTANOU DICAG RAT OF TH ISOLATE COKOC

20. "

,'*N~.j(c) (

AFigure AEFF 'T, O 9(-GBl

" ,()GB1 - 6 M+1%C2 € R 0 4 M(d) GH 0

4'+I~ C2 e lb1

" H N f lB10

3H+t'C2

O.v \H PNAEU XCAG AE FTEIO E OKOC N

G~ O~20

cic

(a)(b)

C*.

°4

AXA- 4.

(c) (d)

....

(e)

Figure 6

EFFECTS OF (a) GHB 5 x 10"3 M, (b) GHB 5 x 10'3 M + 10% CO2, (c) GHB 5 x 10-2 M,

(d) GHB 5 x 10"2 M + 10% CO2., AND (e) SALINE + 10% C02 AFTER GHB TREATMENT,

ON THE SPONTANEOUS DISCHARGE RATE OF THE ISOLATED COCKROACH CNS

21

3. Studies on the Site of Action of C02, GABA, and Related

Compounds

The possibility of using N-acetyl-GABA (NAG) as a chemi-cally stable analogue of GABA-C02 was,previously discussed,and this compound was shown to be stimulatory (ref. 4). Aseries of representatiVe experiments (Figure 7a-c) show howthe electrical discharge rate of the ganglion is affected asthe concentration of NAG is increased. There is little orno stimulation at an NAG 9oncentration of 10-6 M (Figure 7a).At a concentration of 10 " M NAG (Figure 7b) a definite in-crease in the discharge rate is observed, and a prominentoscillatory pattern appears. At a concentration of 10-1 MNAG (Figure 7c), the preparation discharges at a very highand relatively sustained rate, and again shows a pronouncedoscillatory discharge pattern over and above the increase inthe overall discharge rate. The decline in discharge rateafter about one minute may be due to exhaustion of the pre-paration.

If NAG is indeed a valid analogue of GABA-C02 , then it ispossible to test whether GABA and GABA-C02 compete for thesame site on the synaptic membrane of. the cockroach CNS. SinceGABA-C02 is a very unstable compound under these experimentalconditions, NAG could be substituted for GABA-C02 to observethe effects on the discharge rate of the preparation by varyingthe concentration of NAG in the presence of a fixed amount ofGABA.

The results of-partially completed experiments designed totest these effects are shown in Figure 8 (a-d) and Figure: 9(a-c). Figure 8 shows the results of trea~ing one preparationwith a constant concentration of GABA (10- M) and varyingconcentrations of NAG. At a GABA concentration of 10-4 M andan NAG concentration of 10-4 M, neither inhibition due to GABAnor exgitation due to NAG is seen. The lack of GABA inhibitionat 10" M corresponds with what was observed by treatment withthis same concentration of GABA in a previous experiment(Figure 2d). However, excitation was previously seen with thisconcentration of NAG (Figure 7b). Thus, the stimulatory effectsof NAG are definitely modified in the presence of a concentrationof GABA that is barely sufficient to effectively inhibit the bio-electric activity of the preparation. These results could beexplained if these two substances compete for the same site.

22

(a)

ji

I N~~,m . .'r'j :iiv J

0 MIN

(b)

I ~MIN(c)

I Figure 7

EFFECTS OF VARIlOUS CONCENTRATIONS OF NAG ON THE SPONTAI'EOUS DISCHARGE KATE

OF THE ISOLATED COCKROACH CNS(a) NAG 10"6 N; (1b) NAG 10O4 H; (c) NAG 10"1 H

23

%*G ,34

tn ( .3

aaL ,O'I

u ! " . ."'' * % ' ". '.

* .. . . . ".

(a) (b)

40CO

I I

.... ', ,.

.,c.

() (b)

Figure 8EFFECTS OF ADDING (a) NAG 10-4 M, (b) NAG 10 - 3 M, (c) HAG 5 x 107 3 M

AND (d) NAG i0- 2 H, TO THE COCKROACH PREPARATION DURING TREAThENT WITH 10"" M GABA

24

(b)

II.. ...

j "

°°*

(b)

EFFECTS OF ADDING (a) NAG 103 , (b) NAG 5 x 103H, AND (c) NAG 102H,TO THE COCKROACH PREPARATION DURING TREATMENT WJITH 10' M GABA

25

Figure 8b shows the effects of 10 "3 M NAG in the presenceof 10 M GABA. With this mixture there does appear to besome stimulation when NAG is added. In the mixture that con-tains 10-4 M GABA and NAG at 5 x 10 - 3 M, a stimulation with anoscillatory discharge pattern becomes clearly evident (Figure8c) and as the NAG concentration is further increased to 10-2 M

in the presence of 10-4 M GABA (Figure 8d), the stimulation andoscillatory discharge patterns are further increased and sus-tained.

In order to further test these effects, the experiment wasrepeated using a constant GABA concentration of 10-3 M. Thisamount of GABA was previously shown to have a definite inhibitoryeffect on the discbarge rate of the preparation (Figure 3c). Itcan be seen in Figure 9 (a-c) that treatment of another prepara-tion with 10-3 M GABA definitely decreases the discharge rateof the preparation in all cases. Addition of NAG at a concen-tration of 10-3 M barely affects this inhibition (Figure 9a).A greater stimulation is seen when a concentration of 5 x 10 - 3 MNAG is adCed to the GABA solution (Figure 9b), and the prepara-tion is highly §timulated after a lag of about 1.5 min with theaddition of 10- M NAG to the GABA (Figure 9c), and an oscil-latory discharge pattern is again discernible.

Though other experiments remain to be done in this series,some interesting observations can be made. It appears that NAGand GABA do indeed compete for the same site, since increasingthe G^3A concentration delays the time of onset for the expres-sion of the stimulatory effects of NAG. The concentration de-pendence of these effects, together with the fact that a non-in'ibitory concentration of GABA seems to decrease the sen-sitivi-y )f the ganglion to NAG, indicates that these twosu',h;tances are competing for the same site. If NAG is a validanalogue of GABA-C02 , then it follows that GABA-C02 also actson the same receptor site. If this is indeed the case, thena molecule of GABA, a neuroinhibitor on the receptor site,can in Lhe presence of CO2 immediately become a neu:otransmitterby forming the carbamino compound at the same receptor site.The only rate-limiting step controlling this respo se is thediffusion rate of CO2 in the tissue, and not that of the largerand more polar molecule of GABA-CO2 . Therefore the responseto increased CO2 tension in specialized insect nervous tissueis directly controlled by the partial pressure of C02,and almostimmed-ately expressed.

26

From data on the rate constants of the reaction of beta-alanine with CO2 (ref. 10) it can be calculated that thehalf-life of the carbamino compound is about 140 milliseconds.Beta-alanine is structurally very similar to GABA, differingfrom GABA by only one carbon atm between the carboxyl andamino groups. If the half-life of the carbamino-GABA com-pound is of the same order of magnitude as that of carbamino-beta-alanine, then it seems clear that the carbamino-GABAccmpound will be extremely responsive to C02 -tension, andwill exist only as long as the partial pressure of CO2 issufficiently high for its formation in effective quantities.it will spontaneously and rapidly decompose when the partialpressure of C02 decreases.

Thus, an automatic and highly sensitive neurochemicalmechanism is available to the insect for the control of C02tension. Increased nervous activity in specialized receptorstructures due to increased CO2 concentration and carbamino-GABA formation could ultimately lead to the initiation ofappropriate responses to control CO2 tension. Such receptorscould, for instance, be advantageously situated in the neuralapparatus for the control of insect ventilation. In this re-gard, it is of interest to note again the oscillatory patternspreviously mentioned that are exhibited by the preparationwhen treated with NAG and the NAG-GABA-mixtures. It is pos-sible that these oscillations are actually reflections of anrntrinsic "breathing" or ventilatory neural discharge patternthat is triggered by the various treatments.

It has long been known that respiratory movements in the:-sect are under both nervous and chemical control (ref. 11).kespiratory bursts in the isolated nerve cord of the prayingmantis occur with the same frequency as the normal restingventilatory movements of the intact insect (ref. 12).

The actions of the spiracles are coordinated by the centralnervous system in P. americana. CO2 affects the respiratorysystem in this inse-ct and other cockroaches by a direct actionon the ventral nerve cord, which results in entilatory move-ments (ref. 13). Guthrie and Tyndall (ref. 13) state, 'The pre-sence of sensory structures in the spiracles, responsive torelatively low tensions of C02 , suggests a useful local controlmechanism, in addition to the central one in the nervous system,which could be used to select appropriate ventilation flow."

27

If it is ,.stablished in principle that carbamino-GABAis a stimulatory neurotransmitter as we have postulated,than the formation of this substance as a direct functionof CO concentration has obvious relevence to the funda-mental mechanisms governing the nervous control of ventilationin insects. The results of these inyestigations also supportthe hypothesis that originally motivated this work, i.e., thatthe formation of GABA-CO in specialized neural receptors ofthe mosquito is the mechinism whereby increased CO- tensionstimulates mosquito activity and triggers host-seeting be-havior (ref. 5).

'It will be our immediate goal in future investigationsto test the effects cn the preparation of mixtures of NAGwith higher and lower GABA concentrations in order to fur-ther observe these trends. We will also observe the effectsof mixtures of GHB with NAG which could yield further in-formation regarding the site of action of these substancesand relative affinities for the site. Furthermore, sinceKerkut et PI (ref. 14) postulated that glutamate may actuallybe the stirrilatory transmitter in c ockroach myoneural junc-tions, we will test the effects of glutamate, and mixturesof glutamate and GABA. Glutamate closely resembles GABA-C02structurally (ref. 4), and it will therefore be of interestto determine whether glutamate and GABA-C02 act at the samesite.

III. STATISTICAL TESTS OF "BITOMETER" METHOD

Our basic assumptions in applying the "bitometer" methodto test mosquito repellents are:

a. Factors that affect the biting rate of mosquitoeson untreated control mice will also effect thebiting rate of mosquitoes on repellent treatedmice.

b. There are no differences between the data outputof the various meters used in these tests.

c. The utilization of the electronic biting data,i.e., the amount of actual biting time as elec-tronically recorded (P), makes an independentand significant contribution to total assay ofrepellency, regardless of percent mosquitoengorgement (E).

28

Some support for the validity of the last assumption wasobtained when it was revealed in a discriminant functionanalyses (ref. 2) that the parameter for the percentage ofactual biting time (P) makes a significant contribution tothe total assay of repellency. Therefore, this parameterwas used to determine the repellency index by adding P tothat of the percentage of engorged mosquitoes (E). However,we have never tested P independently to analyze interaction.The other two assumptions were never tested.

In order to further document cur repellency testing method,and in view of the observations by Gouck and Smith (ref. 15)that mosquito age and time of day influences the biting rateof low-level repellent treated subjects, we designed a compre-hensive statistical plan to test this effect, as well as someother possible effects in our repellency assay method. Thisexperimental design allowed for independent analysis of thefollowing.

1. A reevaluation of our previous findings ofday-to-day variations in control tests (dayeffect).

2. A comparison with the result of Gouck andSmith (ref. 15) showing the effect of mos-quito age and time of day on avidity in thepresence of low-level repellent treatment.

3. The consistency of the responses among thethree "bitometer-timers" we use in ourtesting procedures.

4. The influence of the actual number of mos-quitoes in the test upon the test results.

5. The possibility of variations among differentbatches of mosquitoes.

The test was arranged with 14 days of testing. Each testday consisted of 4 sets, 2 in the morning, and 2 in the after-noon. Each set utilized 2 repellent treated mice and 1 untreatedcontrol. Each set also contained 2 different ages of mosquitoesrun in parallel at different times of day on the 3 availablemeters.

29

Sixteen separate batches of mosquitoes were utilized toprovide mosquitoes of ages 2, 4, 6, 8, 10, 12, 14, and 16 daysof age in a randomized block design. The complete designcalls for 56 untreated controls and 112 repellent treatedtests, and the results were evaluated for the test and controlobservations. This experiment was designed to supply a compre-hensive basis for future evaluations or applications to massscreening of repellents.

The design of the test is shown in Table I. It is a ran-domized block design, and each mosquito age is tested at leastonce with each meter and at different times of day. Breed'ngof mosquitoes was carefully controlled so that emergence o(-curred at the proper time to match the age schedule. Lightingwas automatically controlled to g've 11 hours light and 11 hoursdarkness, with one hour for slow dimming and one hour for slowbrightening of the lights to simulate dusk and dawn.

Dimethyl phthalate was used as the test repellent. Onesquare inch of the belly of the test mice was treated with0.5 cc of acetone containing 0.05 mg of repellent, while con-trol mice were similarly treated with acetone alone. Onlythe treated areas of the mice were exposed to the mosquitoes.The test methods were as previously described (ref. 16,17).

The summary table (Table 2) shows the mean and 95% confid-ence limits for the control and the test observations for thevariable age and time. Mosquitoes aged six days or less showeda lower level of response in the controls than more matureadults. The response in the treated group was consistent withthis pattern, but did not show enough variation to be signifi-cant. Figure 10 shows this age effect graphically.

Table 2 also shows the repellency index, the sum of P plusE, for tests done at different times of day. The controlsshowed repellency index increasing throughout the day. Inthe repellent treated test group, the response is very lowwith the first morning test. It is lower than the otherthree (the second morning test and both afternoon tests).The variation in the other three tests is not statisticallysignificant. Figure 11 graphs the effects of time of dayfor control and repellent treated cases.

30

I 4J

4i I4(n 4 UI. C'DUC Nt0

440

I 7. 1 U 00UC4UID4C4I C

r-4 *1-4

E--~

HI -)40 1-)h r- -4 . r- 4 r--r--4-

p 41

U) M S4

C) 0- 0

'~14J4 0.

H r-4 1

H a a)

EZ H Ie1 0CNO ~ '0 ~C.~

4-1 04~. 41-1-

zH u N U4N D0 N0QN' uU - r

u. z - - 4r- 4lr40 > cd

131

Table 2

REPELLENCY INDICES WITH STATISTICAL CONFIDENCE LIMITSFOR AGE AND TIME OF DAY EFFECTS

IN CONTROL AND REPELLENT-TREATED CASES

Controls Repellent TreatedAge Mean 95% C.L. Mean 957, C.L.

2c a2 73.4 55.7-91.2 2.0 0-10.44 53 .4ab 35.7-71.2 9 .2ab 0.8-17.6

6 30.3a 12.6-48.0 3.8 a 0-12.2

8 9 8 .3cd 80.6-116.0 11 .2ab 2.8-19.6

10 8 3 .8cd 66.1-101.5 1 7 .5b 9.1-25.9

12 99.0 d 81.3-116.7 11.9 a b 3.5-20.3

14 102 .4d 84.7-120.1 16 .9b 8.5-25.3

16 85.2cd 67.5-102.9 8 .3ab 0-16.7

Time

1st AM 6 1 .9a 49.4-74.4 2.7d 0-8.6

2nd AM 71.2 a b 58.7-83.7 1 1.1 b 5.2-17.0

ist PM 81.8 b c 69.3-94.3 15.4 9.5-21.3

2nd PM 97.8 c 85.3-110.3 11.2 b 5.3-17.2

Superscripts a, b, c, and d represent statisticalgroups. Overlapping groups are denoted by morethan one superscript in the mean values.

32

120

i

\ I

80 \CONTROL

\95 56 LIMIT

01

\ II

z

\ I-J40- \wLU

20-

L.

10\ I 1lS • % !

20 -' I-- ,-, ..,..', "95 %' LIMIT

o.- " ". -" .la ,. -./. .. ...... \TREATED

S 2 4 6 8 I0) 12 14 16

MOSQUITO AGE (DAYS)

Figure 10

EFFECT OF MOSQUITO AGE ON THE REPELLENCY INDEXOF CONTROL AND LOW-LEVEL REPELILENT-TREATED CASES

120-

/95 b LIMIT

//I00 /

100 / / CONTROL

80 959 6 /L IMIT

80 // / /

X

w -

0 60

z-.JU -,I

-J 4 0wLiJ

20 95 LIMIT

-s ° " TREATED

'---9596 LIMIT

0 2 34ISTAM) (20AO (ISTPM) (2NDPM)

TEST NUMBERFigure 11

EFFECT OF TIME OF DAY ON THE REPELLENCY INDEXOF CONTROL AND LOW-LEVEL REPELLENT-TREATED CASES

314

The confidence limits and the tests of significance wereba-ed upon the analysis of variance. Table 3 shows the analysisof variance of the repellency index for the controls. The dayeffect and age effect are significant at the 0.001 level, whichmeans that this result could be the result of random variationwith probability less than one-tenth of one percent, (i.e. itis probably not random). The effect for time of day of theexperiment is also significant (0.005).

Common variation is not specifically associated with in-dividual variables, but is accounted for by more than onevariable. When it is negative the common variation requiresthat the variables are analyzed jointly to account for partof the sums of squares. Common variation can result from arandomized experimental design where some of the variablesmay not be completely balanced. When more variables than oneare required to account for a portion of the sums of squaresin an analysis of variance, then negative values of the commonsums of squares can occur. These can be caused by random effectsand con be safely ignored when they do not change the signifi-cance of variables in the analysis.

Table 4 shows the analysis of variance of the repellencyindex for the repellent treated test cases. The day effectand the time effect were significant at the 0.05 level. Allother variables were not significant (n.s.).

No difference between meters was significant for eithercontrols or test cases. The effect of accidental variations-n the number of mosquitoes was tested and was not signifi-cant.

Although the repellency index, derived from disciminantfunction analysis, is the sum of the percent of actual bitingtime as electronically indicated (P) plus the percent mos-quitoes engorged (E), it was also of interest to analyze thesevariables separately for the control cases. Comparison ofTables 5 and 6 shows that engorgement (Table 5) has the samesignificant variables as percent biting time (Table 6) but atlower levels of significance. Tables 5 and 6 show that thevariables day, age, and time are significant for both P andfor E.

35

Table 3

ANALYSIS OF VARIANCE OF REPELLENCY INDEX FOR CONTROLS

Sums of MeanSource d.f. Squares Square F P

Day 13 29,649.1 2,280.7 4.340 <.001

Age 7 19,728.7 2,818.4 5.360 <.001

Meter 2 1,063.0 531.5 1.010 n.s.

Time 3 9,291.3 3,097.1 5.885 <.005

No. of Mosquitoes 1 1.4 1.4 0.003 n.s.

Error 29 15,253.5 526.0

Common - 1,610.3

Total 55 76,597.3

Table 4

ANALYSIS OF VARIANCE OF REPELLENCY INDEXFOR REPELLENT-TREATED CASES

Sins of Mean

Source d.f. Squares Square F P

Day 13 5,670.4 436.2 1.845 <.05

Age 7 2,505.3 357.9 1.509 n.s.

Meter 2 544.8 272.4 1.152 n.s.

Time 3 2,323.3 741.1 3.139 <.05

No. of Mosquitoes 1 252.2 252.2 1.066 n.s.

Error 85 20,099.6 236.5

Common - -222.7

Total 111 31,172.9

36

Table 5

ANALYSIS OF VARIANCE FOR ENGORGEMENT (E) OF CONTROLS

Sums of MeanSource d.f. Squares Square F P

Day 13 3,967.6 305.20 1.812 < .10

Age 7 3,378.5 482.65 2.890 <.025

Meter 2 204.4 102.18 0.612 n.s.

Time 3 1,499.4 499.79 2.990 <.05

No. of Mosquitoes 1 54.4 54.36 0.325 n.s.

Error 29 4,842.0 166.96

Total 55 13,946.2

Table 6

ANALYSIS OF VARIANCE FOR PERCENT BITING TIME (P) OF CONTROLS

Sums of MeanSource d.f. Squares Square F P

Day 13 11,948.9 921.91 3.030 < .01

Age 7 10,477.1 1,496.93 4.923 < .005Meter 2 1,303.7 651.85 2.142 n.s.

Time 3 3,555.4 1,185.13 3.898 <.025

No. of Mosquitoes 1 38.7 38.70 0.127 n.s.

Engorgement (E) 1 155.2 155.20 0.510 n.s.

Error 28 8,522.2 304.36

Total 55 36,037.2

37

Table 6 also shows that after all these other variablesare Laken into account, the variable percent engorgement (E)is tested. The results show that the variables P and E arenot significantly correlated except through the effect of theother variables. Thus, these two variables prove to give in-dependent information which helps to provide more sensitivetests of the effect of repellent treatments.

Gouck and Smith (ref. 15) found that avidity increasedrapidly with age for the first 5 or 6 days, then remainedfairly uniform. They also found that the morning avidity wasmuch lower than during the previous afternoon,despite theadditional age. Our results (Figure 10) show that there isa continual decrease in avidity up to about 6 days of age,and then a rapid increase after day 6. The same patternappears with both control and repellent treated cases. Ourresults, however, agree with those of Gouck and Smith in ourrespective findings of variations of avidity with time of day.There is a decrease in avidity in the first morning test whichis not significantly different from the second morning test(Table 2), but is highly significantly different from theafternoon tests. The second morning test is not significantlydifferent from the first afternoon test. There appears to bea. decrease in avidity in the second afternoon test in the re-pellent treated cases, and an increase in avidity during thesame test period in controls (Figure 11). However, Table 2shows that there is actually no significant differences foreither the controls or the repellent-treated cases in theset'o time periods, since they fall into the same statistical groups.

In conclusion, the results of these tests appear to sub-stantiate the validity of all the assumptions implicit in thistest methoO. The biting rate of mosquitoes on untreated con-trols parallels that of repellent-treated cases; there are nosignificant differences between the meters; the contributionof the parameter P is independent of that of E, and thereforeprovides unique information which contributes to the sensitivityof the overall repellent testing method.

38

IV. REPELLENCY ASSAY OF SELECTED COMPOUNDS

During this year a number of commercially available com-pounds were selected to be tested for repellency by the elec-tronic recording method, together with the known repellentsDEET and DMP for comparison. The compound names, computerlisting, and structural formulas are as follows:

COMPOUNDCOMPOUND NAME LISTING STRUCTURE

.n-Diethylt.Luampide DEET CH3

8-N(C2H5)2

Dimethyl phLhalate DMP 0

" C-O-CH3

'I C-O-CH3

• ,, thranii A14451-7

-N

Diethyl A14322-7 0(Dimethylaminomethylene) - 0

malonate C-CH C-OCH

CH3 C-OC2H5III0

39

COMPUTER

COMPOUND NAME LISTING STRUCTURE

Diethy! A14390-1 0Dimethylmalonate C-OC2H5

CH3 - -CH3

C-OC2 H5

1 OC 11

0

N,N-DY*nethylfornamide A14277-8 CROC1idiethyl acetal 2 5~r

N,N-Dimethylformamide A14275-1 CH3 0--jethylene acetalcH N I

C3

Ethyl-4-n-propyl-l- A14154-2 0piparazine-carboxylate 11

C-OC2 H5

N

ICH2CH2CH3

2-Methoxyethylamine A14369-3 CH3-O-CH 2CH2NH2

40

COMPUTER

COMPOUND NAME LISTING STRUCTURE

3-Penten-2-one A14501-7 0

CH3-CH-CH-t!-CH 3

4-Phenylbutylamine A14539-4 ( CH2CH2CH2CH2NH2

2-Acetamido-3-butanone A280-7 CHC=O

NH

CH3 -CH

C="OI

cHt3

The computer outputs of the results of the repellency assaysof these compounds are shown in the Appendix. It should be re-called that the number labeled "upper bound" gives the confidencelimit for repellency of the compounds. Numbers below 100 forthe upper bound are significantly repellent at the 95% level ofconfidence, numbers above 100 are not significantly repellentat this level of confidence. Relative merit is indicated bythe magnitude of the upper bound. A compound with a low numberindicates that it has better proof of repellency than a compoundwith a high number, ahtough both may be significantly repellentat the 95% level if both upper bounds are less than 100 (ref. 16).

41

It can be seen that both DEET and DMP are significantlyrepellent at an application level of 0.01 mg/sq in. of mouseskin. DEET may retain borderline repellency at a concentra-tion of 0.001 mg/sq in., but DMP is not significantly repel-lent at this level.

It was previously reported that amino alcohols, aminoaldehydes, and other compounds that appear to be related tothe structure of GABA in terms of electronic distributions,i.e. having at least one nucleophilic center and one electro-philic center in the same molecule, usually show repellencyat low treatment levels (ref. 2,3,4). Compounds that containboth nitrogen and oxygen atoms in certain relationships withinthe same molecule exhibit repellency. Also, compounds thatcontain a double bond between 2 C-atoms instead of aminonitrogen also provide a nucleophilic center, and show repel-lency when an electrophile is also present in the same mole-cule (ref. 3). The following compounds were tested to gainfurther insight into the chemical nature of a repellent.

Anthranil (computer listing A14451-7) showed no repellencyat a concentration of 1.0 mg/sq in., and was not tested further.It is apparently detrimental to repellency to have both thenitrogen and oxygen atoms within a ring structure in whichresonance and charge "smearing" can occur.

Diethyl (dimethylaminomethylene)-malonate (computer listingA14322-7) is significantly repellent at all 3 levels tested(1.0, 0.1, and 0.01 mg/sq in.). Here, two electron-releasing(nucleophilic) sites in the amino groups and the double bondedcarbon atoms are complemented by two electron-withdrawing(electrophilic) groups in the dietbyl malonate moieties, andthe molecule exhibits significant repellency at the levels tested.

Diethyl dimethylmalonate (computer listing A14390-1) is notsignificantly repellent at either of the test levels observed(1.0 and 0.1 mg/sq in.). In this molecule, the electron-with-drawing (electrophilic) ester groups are not complemented byany electron-releasing groups, and the molecule is not repel-lent.

N,N-dimethylformamide diethyl acetal (computer listingA14277-8) shows significant repellency at both 1.0 and 0.1 mg/sqin. We have previously shown that amino acetals are fairlygood repellents (ref. 2,3).

42

N,N-dimethylformamide ethylene acetal (computer listingA14275-l) shows no repellency at either the 1.0 mg/sq in. orthe 0.1 mg/sq in. level. Apparently, when the oxygen atomsare in a saturated ring structure, it is very detrimental tothe repellent properties of the molecule.

We next tested the repellency of ethyl 4-n-propyl-l-piperazine-carboxylate (computer listing A14154-2). Inthis molecule, 2 nitrogen atoms are in a 6-membered saturatedring, and 2 oxygen atoms are present in an ester linkage. Wepreviously discussed the fact that when oxygen atoms were to-gether in a molecule in an ester or an acetal linkage, repel-lency was retained. These findings were interpretated in termsof their similarities and dissimilarities to the structure ofuABA, which contains one strongly electrophilic group (thecarboxyl group) and one strongly nucleophilic group (the aminogroup). The repellency of a substance could be shown if thenitrogen atom was surrounded by carbon atoms, but not theoxygen atom (ref. 2,3). In the compound ethyl 4-n-propyl-l-piperazine-carboxylate, nitrogen atoms are surrounded bycarbon atoms in a ring structure, and two oxygen atoms arepresent in an ester linkage. This compound is significantlyrepellent at the 1.0 and 0.1 mg/sq in. lcvels. Thus, althoughthis compound is totally different structurally from many otherrepellent compounds we have tested, comparison of the testresults of this compound with that of N,N-dimethyl formamideethylene acetal (A14275-1) seems to bear out the basic principlethat modification of the electrophilic properties of oxygen byinsertion into a carbon chain will significantly reduce re-pellency, while the nucleophilic properties of nitrogen arerot greatly modified by other electron-releasing groups, sincethe amino group is itself electron-releasing.

An interesting possible departure from this rule is seenin the compound 2-methoxyethylamine (computer listing A14369-3).Here, the oxygen is surrounded by carbon atoms, but it is inthe form of a methoxy ether-linkage. This is the first aminoether we have tested. This compound proved to be significantlyrepellent at all levels tested (1.0, 0.1, 0.01 mg/sq in.). Itwill be of interest to further investigate the repellent pro-perties of compounds with this type of structure.

43

We previcus h' reportcd t"ha:: 34but ene-2-o I was sign i fcan tiyrepellent at an applica:-vmo lcvei. 'zf .0] mag/sq in. (ref. 3).In this series of tests, we tested the repellency of 3-penten-2-one (computer listing A14501-7). This compound showed no re-pellency at an application level of 1.0 mg/sq in. and was nottested further. In this compnund there is the possibility ofresonanice between the unsaturated doubic bonded carbon atomsand the keto-group. Such resonance would tend to "smear" outthe charge difference between these two groups, and thus vitiatethe charge separation between the n-ucleophile and electrophile(i.e., between the do.uble-bded -C=C- group and the keto-grouprespectively). Under these conditions, repellency is apparentlycompletely lost. Structurally the c,.;mpounds 3-butene-2-ol and3-penten-2-one are not very dissimilar. This finding againtends to substantiate the hyp.'thesis that the electronic dis-tribution in a molecule rather than its actual atomic con-stituents is the main factor chac determines the repellencyof a compound.

4-Phenylbuty lamire (c"'.-puter listing A14539-4) was nexttested for repellencv. This cr.mnpiund was fo-cnd to be signi-ficantly repellent at concentration's of 1.0 and 0.1. mg/sq in.but not at tower levels. Tho-ugh the existence of the nucleo-phile (the amino group) is ev~ident, this compound cannot easilybe shown to have compl-mentary nucleophilic and elect~rophiliccenters. This 4-carbon c-_-:npo:und, hcwever, may have a stericresemblence to the structure of GABA. The e-.ident and unex-pe.1ted repei'ilencv -.f 'this .:m'uddcmn.istrates that therenav exist more than one ine:.har'ism b - which a substance mayC'.use avoidance react . .s .in a m-squt'

The last comnpound, 2-acetamido-3-butanone (cOmlputer listingA280-7) is of spec2 Ll interest. In view -f the physiologicalactit~n of N-ce.v GB (pre-.vus l.v shown in the coc.kroachpreparation), it was of irtecest b) Lest. a %vo]atile analogueof NAG for repellency, ramely N-acetyl.-4-butanol (NAB). Uponconsulting chemical catal.:gues and chemical literature, wefound that this ccompnund is roL comnercially available, andindeed, has never been reported Lo 1-he literature. We there-fore chose 2-acetamido-3-butanore to vest for repellency, sinceit seemed to come closest to the desired structure and is corn-mercially available. Althouigh 2-acetamido-3-butanone is aketone, and we previously found ke'tones not to ba very repellent,this compound retained significant repellency to a concentrationof 0.1 mng/sq in. Though it was n-'t repellent below this level,these results are very enc-ouragirg. If possible, we will, at-tempt in future work to synthesize NAB to test for repellency.If it is repellent, the mechanism by which it exerts repellency

44

willsee clar.By anaog~to. the highiv stimul tory effects%that we found NLAG to exert in' the cockroach central nervous

*syster., NAB would increasingly stimulate the neural receptorstructures of the mosquito as a Prospective host is approached,and thus unibalance anJ confuse thte d-elicate mechanism governing

*%the hcsc-f--ndin- activities of the ".squito.

V. SUM2ARY

A.Electrophvsiologica. Investizatioals

Ouring tcbis year detailed sy-sematic studies were conductedco test the hypothesis that gari-a-ainobutyric acid (GARA), aneuro~nibitory substance that is found in mosquito extracts,Duecomes neuroeveitatorv ;.hen it reacts with carbon diioxide toforr a narbamiro compound. It was sho;wn that treatment of theisolated cockroach central nervous system with GAMA reduces itsrate ef sp.)raneous bioelectric discharge. When the preparationis treatec: .-ith C01, alone, or with a GARA solutior that hadA.10! "'02 bubhble: inlo it, the spontaneous discharge rate in-creases.

in order to- distinguish between the stimujlatory action ofCO-) ali ne and carbanino-GAEA, the effect of ga~na-hydroxybutvricacid (GHB) was observed in the Presence and absence of C02. GH".annot form a carbamino compourd since it does not contain anamino group. It was found that in the presence of GFB the sen-Citivity of th2 preparation to CO? is completelv abolished.This result implies that carbamin5-GABA formation is the only

nehnimb wihCO0) acts as a neurostimulator in this pre-paration. Astable chemical analogue of carbamino-GABA, N-acetyl GABA (NXAG) was applied to the preparation, and found tobe neurostixnulatorv. Cox.petition stud~es between NAG and CAMAshowed that both these substances appear to compete for thesame site on the synaptic membrane. This observation providesadditional evidence that carba-nino-GABA formation, is the mecha-nism by which C02 stimulates the insect centrci! nervous system,and does not act through a separate receptor site for 0102.These results were related to the possible mechanism by wihichC02 exerts central and local control of ventilation in insects,and triggers the host-seeking activities of mosquitoes.

45

B. Studies of Repellency Assay by the Electronic Method

A statistically designed series of tests were performedto test the assumptiors that (1) variables affecting thebiting rate of mosquitoes on untreated control mice will alsoaffect the repellent-treated cases in a parallel way; (2)there are no significant differences in the data output amongthe three meters used in the electronic recording method; (3)the electronically recorded data makes a significant and in-dependent contribution to the total assay of repellency. Thesethree assumptions were statistically proven to be correct, andthe validity of the electronic repellency assay method was thusfurther documented.

C. Tests of Repellents

A number of compounds were tested for repellency that wereselected to further test the rationale that if a substance isvolatile, and has an electronic structure similar to that ofIABA, i.e., contains at least one nucleophilic and one elec-trophilic center, it will be repellent. Though an exceptionwas noted, it was shown that in general substances that hadthis electronic structure were repellent, while substancesthat did not have this structure were not repellent. Thisfurther substantiated previous findings. The possibilitywas discussed that volatile chemical analogues of anotherph-siologicallv active substance discovered in the courseof this work, N-acetyl. GABA, may also be considerably repel-ent to mosquitoes.

VI. FUTURE INVESTIGATIONS

The approaches described above and detailed in past reportswill as far as possible be continued in future work. We anti-cipate that the future work program will encompass the followingareas of investigation.

A. Further Investigations of the GABA Hypothesis

Work will continue in the general avenues of research des-cribed in this report. We will further investigate the identityor non-identity of the sites of action of GABA, GABA-C02 , GHB,and glutamate in the insect central nervous system. The en-couraging positive results thus far obtained with the cockroachpreparation will be further explored and substantiated.

46

g

B. Search for New Repellents

The search for new repellents based upon specific require-ments in terms of molecular and electronic configurations ofcandidate -ompounds will continue. Commercially availablematerials w:iil be obtained wheneve.: possible to continue totest these approaches. If possible, we "will attempt to syn-thesize the alcohol analogue of N-aceUyl GABA. The lattersubstance shows Fotenz physiological activity in the insectcentral nervous s-stem. This alcohol analogue, 4-aceramido-butanol, ria% e.xhibit substantial repellency.

M?ntenance of Mosquito and Cockroach Colonies

Our cdoldies of Aedes aegvpti (L) mosquitoes and erpitaarnericana (L) cockroaches will continue to be propagated andmaintainied for purposes of our research throughout the courseof this work.

VI i. CON'LUSIONS

Our investigations to date have included 3 main areas ofresearch:

1. Testing a hypothesis that could explain thebasic physiological mechanisms that drivem3squitoes to warm-blooded hoses.

2. Developing accurate statistically-basedmethods to evaluate mosquito repellency oftest comp3unds. These methods employ a de-vice that electronically records, and thusunequivocally documerts the mosquito bite.

3. Testing the mosquito repellency of certainchemical structures that are rel. Led to thestructure of GABA and exploring the natureof a repellent and the mechanism of repel-lencv. This work is directed toward obtaininga rationale to guide the discovery of morepotent and longer-lasting repellents thanthose currently available.

47

We plan to continue using the unique theoretical andmethodological approaches developed during the course ofthis work. The attempt to elucidate the basic ,)hysiologi-cal mechanisms governing the interactions of mosquitoeswith their warm-blooded hosts, and the logical approachesto the development of new insect repellents engenderedduring this work are bearing fruitful results. Continu-ation of this effort should substantially contribute toultimate success in the important endeavor to achievesuperior repellents for systemic or topical use.

48

I

LITERATURE CITED

1. Kashin, P., "Development of an Orally Effective InsectRepellent," Annual ProgresE Report No. IITRI-L6021-4,Contract No. DA-49-193-MD-2281 conducted by IIT ResearchInstitute, Chicago, Illinois, October 31, 1965.

2. Kashin, P., "Development of an Orally Effective InsectRepellent," Annual Progress Report No. IITRI-L6021-8,Contract No. DA--9-193-MD-2281 conducted by lIT ResearchInstitute, Chicago, Illinois, November 11, 1966.

3. Kashin, P., "Development of an Orally Effective InsectRepellent," Annual Progress Report No. IITRI-L6021-12,Contract No. DA-49-193- -- 2281 conducted by lIT ResearchInstitute, Chicago, Illinois, November 1967.

4. Kashin, P., "Development ef an Orally Effective InsectRepellent," Annual Progress Report No. IITRI-L6021-16,Contract No. DA-49-l93-,D-2:81 conducted by ITT ResearchInstitute, Chicago, Illinois, November 1968.

5. Kashin, P., Ann. Entomo]. Soc. of Amer. 62, 695 (1969).

6. Smit, W. A., Becht, G., and Beenakkers, A. M. Th., J.Insect. Physiol. 13, 1857 (1967).

7. Edwards, C. and Kuffler, S. W., J. Neurochem. 4, 19 (1959).

8. Margolis, R. K., Biochem. Pharmacol. 18, 1243 (1969).

9. Kashin, P., Quarterly Progress Report, IITRI-L6021-17,January 31, 1969.

10. Jensen, A. and Faurhalt, C., Acta Chem. Scand. 6, 385,(1952).

11. Herford, G. M., J. Exptl. Biol. 15, 327 (1938).

12. Edwards, G. A., Respiratory Mechanisms, in Roeder, Y. P.,Insect Physiology, John Wiley-Son, Publ., New York (1953).

13. Guthrie, D. M. and Tindall, A. R., Respiration (Chapt. 16),in, The Biology of the Cockroach, St. Martin's Press, Publ.New York (1968).

49

14. Kerkut, G. A., LeakVIe, L. D., Cowan, S., Shapiro, A.,and Walker, R., J. Comp. Biochem. Physiol. 15, 485 (1965).

15. Gouck, H. K. and Smith, C. N., The Florida Entomologist45, 93 (1963).

16. Kaslk-in, P. and Kardatzkr~e, M. L., J. Econ. Entomol. 62. 10(1969).

17. Kashin, P. and Arneson, B. E., J. Econ. Entomol. 62, 200(1969).

50

A PPENDI X

REPELLENCY ASSAY OF SELECTED COMPOUNDS

N fn t& I- OD ' 0 V) -Z F a-DU 0' 0A fn tO IA -D r- CC 00 C N M 0 zt 44 M L L' (N000

> 44~~ -- -- --- -f,4- N -N -Q4

U-

L-1 I

t C -ADUQQr 'W0 ' M W-WPD~"r -~-~%IU -'rD -n D0,

z

La

u u u.1 qN00 n.;A4J O % 1 -40 N u 'f J u o- .p.. V0 0I' U U Q U U U JU U U L

,U V 0 .J '.LVJ. UCO% UUI( J U~ -Jt 0IUJJ .L~~ .0 ",U J0

.Ic

v w .Pu u w U&j Www w UWUWW% u~ V W W %JV WAW

.U m0 0, 0 - N, 1-~ 4d U w0 a Cj 0 0- N 'o p f- m m' Cdw t- 4 ' 0 t V f' 0 f U, U. w t An U 1 0'. 0 0 0. 43 'a c-

s- 0

4.4

II ia.,'

0 0 LA N

x 0 cc 0 0

0 u .- I 0

C- L I.I

1 0 N 0 F p-4 7 o4 P- CCm ' N 0 4 .- m I" P. m 00

-jx41 L n mr .- IN r0 4 kAm (1 N 4 o

LA I WOj t-fw- a .4 tn -- 0 - 0 -wL.JI w-I04D1 0.z

.9 II 00c NP-Ob0t 09 MaDOLItrO m f O

u I ZE). I r 00' %t - w .P -mamN .U .0U) I D.. CL a. I

X t I

1 I

LUew 1I00 tn 14p-4 n % 0 0. 0 0N 0 -~4 -t M M . .40 6N fn

*) 4 UM m N0

0 'A I I '

a LI

L'i %JVU W UW W JUUUIJUI WJUQWW w

u t QU %~a Q %. v v j u %J w w u w ou w0 j) uv o kCL I-DI uU U U U U .1 .U. oIJw% 0wI Awwv%

31.,

.j I

U, I

u I

A2

64 .CD 1 0c

C) ulN u. NNNN fl v w "M

-a-

-I b-- Z.

-'. N4 c04 r.

r. W, ** * z

oc 0 N 0Dtl

Z 7cc

uj 0 0

0.de 1 0' QD0.

Lo, 0j.1 m'0 a.P- cu

-I ~~j uju -4n 01w I 34.0 I

: I O0 !. 4i~~~

~I L u .

00 0 a4I Go 0 0.- InC4 V~' 00,0 ;A- 'r. I...0 .L-Z. J e3 a=" I'N - P

P- -I I

Ci C

-_j91

LL I 'jOL ~ ~ f

LIIu

w'L u)LI

LI 0 Q 0 00. III1 3

013A3

0L C% x V 0 m0C O

. 49 4C4- 4 oor c 0 c0 0 0:00- 0I- Z0- X I-

10 C? I. C 0

It a) Q . 9

'N N~ lN

0 C N

C ao ccO0f 0 0 0

L uj LU wLL CL 0.. m.00 0. a. 0LU .J40

0.-61- 0

0.0

&A 71, 1-4 (1, r % O.4D1 inN OI w , w -T0 ON C

LUI WCj M W:)1 0.1.a I LU-I

i g w FrL, wJ f" NN N 0I -. Nn-~ ~ ~ U

I- WU I

1i 0

C40. -l- 09 -9N'S t A, eq f a%0 COe.N.. **'4wi 4i P # -I f--4 0.004' V ?f~P ',I a .41 P. LsI%, n

;~ a41- 4M" 'A0 0907 ; 97 Onot' .1 :4I~ uI - 4 M Nt~ i.9 NnI' *

At 0-

%am &0 V%'I 0 IV ofI S U

Z~a I

IL I J I

SL x

a I

-JI £A

C- C

c-coo wo M 0

Lr I-a 2f cof

LA W 0 p4-v. of

ii~~~ umlgiN

o c C C

C.) 0

1L. L, i

0DIOD C. f. C

wA U-' wO n n4- I N- r W

* ,W-I N 0 It N m p0 cIUIJ 1 0I' sAW. .

-. xt I 'aN 0C 1,Pl19 o10

* ._. w I -. 0C rS' (14 P.~ 0 O. NW N

01 C21

Qg 0,

el-I-J Ai' M -1%0 N # N *N0d o"N .4' A 0a

011 1- o. .0.- md 04.

oI '0 0. Nf * 0.0 N4 f!OUJ I '' lI V it Oi OZ i Ol

X O

41 dC )I* ~ ~ -- N 5 - Ns.o i4.wwew w . %0 %

4c w % % i ww l

-'w -IUW% oItQ

i j ~~cV) % j4

wi I

"II in p

w w.I W ~ Ir

a.,5

Q) -

z- -CK 5-

.4. Eh -0O 4 f0 ~ cf. LCI- C-

U',)

I.- 04 cf)-t Z .0OZi -4 -4

4 Nl 4A4-*-M

S0 0 0

IL 4) U. WU.L . .)

tC CL

00 00

VS J W - oLA'e '04JU. 7 0 0 i

0.Z

I U.% .m 0 a-- 0-4 NWN'. W O '0! W 0 0 0 0 0 0000 0 e S* 0

I Z!Q4N. I

AAI

UI90.0. 0

(us, f"

aLw

0- t S - 0 0 * 0 0 0 *W 0 0 0 0 0 * 00 sWoZ. , 0~ IAe0 W4 0~V

m-dN~ ~~~~~ W~ %04- W.- W~f' - ~ -

P.5Pft 14 f

@fl5~ .A6

II

4Jn C 0

'~0in. P.C

z~cc44.4

I-

CLC

V* rj to 00F .NO y y MW n PC, Q0- DW.. DP

qj Ca. a,0..0

.J -

M-- - in M'sLW 1 -. P1N~f- 0 *N 0 in b% *' 0

0, d '. ey 040- y N 10UviI . l 'r .fi fm .

IA , I

I U Z Ii ~ M P -. P- O N C -l

WI -C - _I -w-w W #

.0, WO4;-?-v Iv %

pgs*~~d 4. * * *Os S

A70

000

1-4 W -4.9 N N

0' 0- a I-x

0O ~ ~ 4 U-4 P.

r4 DoP 0

z

do 0do

at.U. NP. 6.-

w WS IL 9 06

P-V I-

~~~~i I 0a~ ~ u1N O6 N.. LAN

V) -j N O N .0 Ar.o ita *ea imS/I .wI N 4N P. t 0' ~ 6 ~WjI w l -a N- %oQ. z w I

a-i uj ~6 a e p r o m

sj

v..a10L ~ -J0 "Q C0 mi Cf

X Is

1g0 oO-N OL N04 -4U IA y a0 EUNII- IA F 10Ob N u r *O'M *47 ona

at I o- c0 0 2 w J 0U,1q r ~0 t V 4 4 . 4w 0 %wo jw%1/u VII L

aI U Il V U 3 0C O

i I .

AdI L inImIw Ir qSI ~D I

'I, 7 7 AS

z a1- ar

wu - a a

DO. Zw -

P. N4

fJ%

C

U. afIL V ^C w WU

a. a.

jw rd

- 4-

Ij IO~ p- mOu o UVI 0 ~ : %

6ji d -0

0~~f C- 0 PO M0w

QQ Q *a -4 -lox- IIM O

2-I

9cc ift 000 10vwso%

I FUJ &%9f

kii

u I

4.JI ~ U9D~ D~D % %*A9

w.a

M 0 0ff

-. NNNN w NN r N w

C) - c-oh-

0- 'Oh P. 01-0

40 CtoT Ix

-- i C4U O

-- F- ml k me m I

La aon i f" 00'4.N P.' 0 Of l

0 'A

Vu1.10 00. %M 000 p -A .tA 94% PW I 4 ~ % s .j 0 ~ N ~ 0 ~ 0 0 4G ~ t4 'Ih

w w

QsrsA %il0 00 %0U V V ri 0

IL&

I I I

03 IO

UNCLASSI FIED

Sec rit cl nx(It.11 nnDOCUM ENT CONTRO L D ATA .- R SO D(Secritjy rIA6..i1jg &l#on of fli, hodpr .. I ah'.ls..rt ouI is.ox,,a,. f,.nl .u h al h. o-..I wI,,', flm la,,,,il ret )4 chesitUeM d)

I ORIGIN A TO NG A CTIV ITV (Corpuieile auIAO1J as -. Olo &CR T LAINFIC T1

IIT RESEARCH INSTITUTE Unclassified10 West 35th Street lb04uChicago, Illinois N/Aj REpoRy TITLE

Development of an Orally Effective Insect Repellent

4. DESCRIPTIVE NOTES (1ripe of PI'Pot and Inctualve dotes)

Annual Progress Report,. November 1, 1968 through October 31, 1969S AU THORISI (First natn,. mInddle Iniial. ls name)

Philip Kashin

6. REPORT DATE ?.TOTAL NO. OF PAGES 7b&. NO. of MEPS

COTAC RATDecember 1969 64 I18B.CNRC ORGATNO. Va. ORIGINATOR'$ REPORT NUMBINIS)

DA-49-193-DA-2281 ITIL012b. : ROJECT NO. ITIL012

dac SI. OT14911 RUPOR1 N0191 (Any other P-Jaboeta t mOSY be Ossigned

10 DISTRIBUTION STATIMENT

Distribution of this document is unlimited.

It- SPPLI[AENTRV NOES 1. SPONSORING MAILITARY ACTIVITY

U.S. Army Medical Research and De-velopment Command, Office of TheSurgeon General, Washington, D.C.

13. %OsrACT

.Experimental observations were made to test the validity of the gamma-amino~butyri c acid (GABA)-carbon dioxide (C02) hypothesis, designed toexplain the neurochernical interactions that govern a mosquito's behaviorin the presence of its host. It was postulated that the formation ofcarbamino-GABA underlies the activating effects Of C02 on mosquitoes.The isolated central nervous system of the cockroach, P. amnericana,

was utilized to test the effects of carbamino-GABA and o'E~er compoundson the nervous system of an insect. G3ABA inhibited the spontaneousfiring rate of this preparation in the absence of Co2 , but stimulatedit in its presence. Gamma-yrxyu i acd a ustance which can-not form a carbamino compound, was inhibitory both in the absence andpresence Of C02. A chemical analogue of carbamino-GABA, N-acetyl GABAstimulated the preparation. It was shown that NAG and GABA probablycompete for the same site on the synaptic membrane. Further studiesare in progress.The evaluation of selected compounds for mosquito repellency by the

electronic recording method was continued. Compounds were selected onthe basis of our past approach, i.e. that: volatile chemical analoguesof GABA would affect the mosquito's responses to its host.A statistically designed series of tests were performed to test the

validity of some basic assumptions that are involved in using the elec-tronic method for the assay of repellents. The results of these testslargelx substantiated-our confidence in th amthod.

DD 1Nov..1473 UNCLASSIFIEDSecuritv ctaimdstIfI'n

UNCLASSIFIEDSergrity C"|u.'sfmu uton

- - S

KI'V WORD L111 L141

HOLd ,T NOLJ WT POL

MosquitoRepellentGamma-aminobutyric acidCarbon dioxideNeuroinhibitorActivationElectronicBitometer-TimerRepellency indexComputerized repellency assay

IiI,

I'

%acurily Cls lefv1

I

II

_ _ _ __ _ _ _ mam

-- - -UNCLASIFIES~cuI~y ~nuaf~ca~~'