Vol. 170, No. 6

A Rapid Population Method for Action Spectra Applied toHalobacterium halobium

W. STOECKENIUS,l* E. K. WOLFF,2t AND B. HESS2Cardiovascular Research Institute and Department ofBiochemistry and Biophysics, University of California,

San Francisco, California 94143,1 and Max-Planck-Institut fur Ernahrungsphysiologie, Rheinlanddamm 201, D4600Dortmund 1, Federal Republic of Germany2

Received 5 November 1987/Accepted 30 January 1988

We have developed a simple and rapid technique for measuring the action spectra for phototaxis ofpopulations of microorganisms and applied it to halobacteria. A microscope with a dark-field condenser wasused to illuminate the cell suspension in a sealed chamber with light of wavelength >750 nm; in this region ofthe spectrum, the halobacteria show no phototactic response. A 150-p.m spot of light from a xenon arc lamp,whose wavelength and intensity can be varied, was projected through the objective lens into the center of thedark field. The objective lens imaged this measuring spot through a 780-nm cut-off filter on an aperture in frontof a photomultiplier. The intensity of the scattered 750-nm light, and therefore the photomultiplier current, isproportional to the number of cells in the measuring spot. A third lamp provided background light of variablewavelength and intensity through the dark-field condenser. To minimize secondary effects due to large changesin cell density, we recorded the initial changes in the photomultiplier current over 1 min after the actinic lighthad been switched on. By plotting the rate of change against wavelength, we obtained action spectra after theproper corrections for changes in light intensity with wavelength were applied and saturation effects were

avoided.

Many motile microorganisms can sense changes in lightintensity and will move toward or away from a light source.These photoactivated responses in bacteria are usuallycalled phototaxis, in analogy to chemotaxis, a similar andmuch more extensively investigated mechanism. We shallfollow this convention, even though phototaxis is usedsomewhat differently by photobiologists mainly concernedwith eucaryotic organisms (for a comprehensive review, seereference 7).Action spectra provide the most important criteria for the

initial characterization of photoreceptors (for a recent re-view, see reference 9). Spectra for the light-controlled move-ments of microorganisms may be obtained by observingeither single cells or cell populations. Observation of singlecells is indispensible for an analysis of the photoreactions;however, this observation needs a large number of measure-ments, is likely to introduce bias through selection of cells,and, if fully automated, requires substantial programmingand computer time (8, 20, 21). Complete action spectra aremuch more easily obtained by monitoring changes in celldensity as a function of spatial differences in light intensityand wavelength. We developed the microscopic methoddescribed here to obtain action spectra for the phototacticresponses of halobacteria, which, until now, have beenobtained mainly by observations of single cells (2, 3, 22, 23)or by a very slow macroscopic technique (14). The basictechnique was already used by Engelmann, the discoverer ofbacterial phototaxis (6), who observed the accumulation ofcells in a field of attractant light. An improved recordingphototaxigraph was introduced by Diehn (4). This instru-ment measured the changes in optical density caused by

* Corresponding author.t Present address: Universitat Witten/Herdecke, Institut fur

Technologieentwicklung und Systemanalyse, D-5810 Witten, Fed-eral Republic of Germany.

changes in cell density in the illuminated volume of a cellsuspension. Its measuring beam had a cross section ofseveral square millimeters. We used a microscope andmonitored the changes in cell density when a spot of actiniclight was projected onto a cell suspension by recording theintensity changes of a nonactinic measuring light scatteredfrom this spot.The same equipment can also be used for following the

reactions of single cells; it is commercially available fromseveral manufacturers and requires only minor modifica-tions. Results obtained with this method, which led to thediscovery of a second photoreceptor in Halobacterium ha-lobium, have been published (24). Here we describe theinstrument and its application in detail. The method shouldalso prove valuable for the study of phototactic reactions inother microorganisms.

MATERIALS AND METHODS

Instrument. We used an Orthoplan microscope (E. LeitzGmbH, Wetzlar, Federal Republic of Germany) equippedfor incident light fluorescence microscopy and photometry(Fig. 1). The specimen was illuminated alternatively orsimultaneously by two Leitz 100-W quartz iodine lampsthrough a substage dark-field condenser with an aperturefrom 0.8 to 0.95, which produced an illuminated field of2-mm diameter in the object plane. One of the lampsprovided the measuring light of wavelength >750 nm

through a Schott RG 780 filter; the others served for visualobservation or, with the appropriate interference filters, as asource of constant background illumination.The actinic light source was the 150-W XBO lamp of a

Universal Monochromator Illuminator (Oriel Corp., Strat-ford, Conn.), combined with a 7340 monochromator and a7271 grating (Oriel Corp.). Light from the monochromatorexit slit was transferred via a rectangular (0.7- by 10-mm) to

2790

JOURNAL OF BACTERIOLOGY, June 1988, p. 2790-27950021-9193/88/062790-06$02.00/0Copyright X 1988, American Society for Microbiology

on May 22, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

TECHNIQUE FOR OBTAINING PHOTOTAXIS ACTION SPECTRA

FIG. 1. Scheme of the instrument and light path. Li, Actiniclight source; L2, measuring light source; L3, background light; M,monochromator; Fl, optical UV blocking filter and shutter unit; F2,infrared filter; F3, interference filter for background illumination;PM, photomultiplier.

circular (4-mm), fused silica fiber optic cable to the incidentlight illuminator of the microscope so that the circular cableend replaced the filament of the lamp in the standardhousing. The spectral resolution was 3.0 nm, and the lightintensity was varied by the insertion of neutral densityfilters. The field aperture of the incident-light condenser wasadjusted so that a 170-jim diameter area in the center of thefield of view was uniformly illuminated through the 32/065 P(E. Leitz GmbH) oil immersion objective lens. This measur-

ing spot could therefore be illuminated simultaneously by theactinic light and through the substage dark-field condenserwith >750-nm measuring light and with a beam splitter,simultaneously with actinic background light, if required.The field aperture of the microscope photometer was

adjusted so that all the light from the measuring spot butnone from the surrounding area reached the photomultiplier,which was also protected from the scattered actinic andbackground light by a Schott RG 780 filter. This procedureassured that only the >750-nm measuring light scattered bythe cells in the measuring spot was monitored. The intensityof this light, in a first approximation, is proportional to thenumber of cells in the spot. The photomultiplier currentcould, therefore, be used to record the changes in cellnumber in response to changes of the actinic light.The Leitz Orthoplan microscope is delivered with filters in

the incident-light illuminator, which must be removed. Thestandard half-silvered mirror, which deflects the light into

A B

C DFIG. 2. Aerotaxis of H. halobium. Cells in the dark accumulate

near an air bubble in the measuring chamber and disperse when theoxygen in the bubble is exhausted. (A to D) At 1, 2, 3, and 4 h,respectively, after the chamber had been filled. Dark-field illumina-tion magnification, x 16.

the objective lens, can be used. A more efficient solution isto replace this mirror with a dichroic mirror (cold mirror)with transmission at >700 nm. Because the cold mirror hasa transmission band in the near UV, a second dichroic mirrorwith a transmission limit at shorter wavelength (UV mirror)is required. Both mirrors have to be aligned carefully so thatthey illuminate the same area in the specimen, or theapertures defining the measuring spot have to be realignedeach time the mirrors are switched. The dichroic mirrors arenot available as standard equipment.To obtain large-area dark-field illumination (Fig. 2), we

used an automobile headlight reflector inverted over a pho-tographic ring flashlight. The chamber with the cell suspen-sion (see below) was placed on the opening for the light bulb,and photographs were taken at a magnification of x6.4 witha Polaroid camera equipped with a Zeiss Tessovar lens.Data acquisition and evaluation. The actinic light was

controlled by a Compur electronic shutter and a home-builttimer which also triggered data acquisition via a strip chartrecorder and/or a computer (model 205-A; Nicolet Instru-ment Co., Madison, Wis.). The data were recorded on diskand later transferred to a computer (model 3032; The Perkin-Elmer Corp., Norwalk, Conn.) for further analysis. Only theinitial slope of the scattering change was used (see Results)and analyzed via least-square methods.The calibration of the light flux through the actinic spot for

different wavelengths was obtained by mounting a siliconphotodiode in the position of the specimen chamber (seebelow). The total light flux was thus measured, including alloptically active elements which might influence the experi-ment and the normalization of the data. The relative respon-siveness of the diode over the range from 350 to 700 nm wasobtained by comparing the output current to the light fluxmeasured via a precision light flux probe (KLC model dc

2791VOL. 170, 1988

on May 22, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

2792 STOECKENIUS ET AL.

1010 lightmeter; Karl Lambrecht Co., Chicago, Ill.). Tomeasure the absolute light flux in the microscope, the samelight beam was focused onto the diode and onto the light fluxprobe of the KLC lightmeter through a 590.4-nm interfer-ence filter with half width of 2.6 nm (Schott Glaswerke,Mainz, Federal Republic of Germany).The responses of the cells at all wavelengths were normal-

ized to a reference wavelength (see Results) and correctedfor fluctuations in actinic light intensity.The calibration factors for the different wavelengths were

then used to correct the action spectrum in the program fordata analysis of the Perkin-Elmer 3032 computer, and thecorrected action spectrum was plotted via a standard routinedeveloped at the Max Planck Institute, Dortmund (availablefrom B. Hess on request).The light flux in the actinic spot used in the experiments

under nonsaturating conditions was in the range of 8 x 1011to 1.6 x 1012 photons mm-2* s-1 at 590 nm, i.e., themaximum of the action spectrum. This light intensity iscomparable with or lower than the intensities used by otherworkers (3, 11, 12, 22, 23).

Cell growth conditions. For the experiments reported here,we used the H. halobium mutant Flx3 or strains derivedfrom it to avoid possible interference from other photoactivepigments. These strains lack bacteriorhodopsin and halorho-dopsin but contain sensory rhodopsin(s) (sR); Flx3R alsolacks retinal synthesis (15, 16).

Cells were grown in complex medium at 37 to 40°C underillumination with white light. To maintain high motility,growth in suspension and on swarmplates was alternated(18). For the phototactic experiments, these cells weregrown in 100-ml shake cultures and harvested at the end oftheir logarithmic growth phase (usually after 65 to 90 h).Strain Flx3R showed no phototactic responses, but theresponses were present after 100 ,u of 10-3 to 10-5 M retinalsolution in ethanol had been added to a 100-ml suspensionand the cells had been incubated overnight or for at least 1 hbefore the experiment started.

Preparation of cell suspensions. Usually, cells showed thelargest phototactic responses near the end of their logarith-mic growth phase, but cultures then also contained a vari-able but significant number of large, poorly mobile cells. Toremove these cells, 30 ml of the culture was centrifuged at2,000 x g for 9 min in the swinging bucket rotor of a HeraeusChrist Minifuge 2. The cells from the supernatant wereconcentrated by centrifugation at 2,000 x g for 30 min ontoa cushion of Fluorolube (Hooker Industrial, Niagara Falls,N.Y.), a water-immiscible liquid of high density, on whichthey formed a distinct band. Most of the band was carefullyremoved with a pipette and suspended to a concentration of2 x 109 cells. ml-' in basal salt, i.e., the growth mediumwithout peptone. At this cell concentration, the suspensionrapidly became anaerobic. It has been reported that blockingthe respiratory chain does not affect or even increases thesensitivity of the phototactic response (1). However, 0.2%arginine must be added to the basal salt solution to ensure asufficient energy supply for the cells (5). Aliquots of thesuspension were transferred to 0.1-mm-deep chambers (10by 20 mm) on a microscope slide, and the chambers wereclosed with cover slips and sealed with molten paraffin wax.All preparative steps up to this point were carried out atroom temperature. Temperature control for the experimentswas provided by placing this observation chamber in ahome-built thermostated holder, which fitted onto the micro-scope stage.While filling the chambers, care must be taken to avoid

trapping air; H. halobium is also aerotactic and accumulatesnear air bubbles (Fig. 2). Air bubbles can give rise tolong-lasting, strong base-line drifts or large changes in baseline when the specimen is moved. Qualitatively, we ob-served the following behavior. After 1 h, a faint ring ofaccumulated cells was seen around the air bubble (Fig. 2a).This ring slowly became narrower as the cell density in-creased (Fig. 2b and c). After about 4 h, the accumulation ofcells disappeared, presumably because the oxygen in the airbubble had been used up. This observation agreed with the-20% reduction of the air bubble volume. We have, on otheroccasions, observed that the cells first aggregated in a ring ofhigher density some distance from the air bubble and thatthis ring slowly moved toward the bubble until the ringtouched the bubble. The determination of the causes forthese variations in behavior will require further work underbetter-defined conditions, especially temperature control.

RESULTS

Halobacteria are attracted by long-wavelength and visiblelight and repelled by short-wavelength, visible, and near-UVlight (3, 18). The resulting accumulation or depletion of cellsin the measuring spot can be observed directly (Fig. 3A andB) or recorded as the change in photomultiplier current (Fig.3C). At constant light intensity, the change in photomulti-plier current is linear for approximately 1 min and thenslowly decreases. The decreasing rate of change is at firstmainly due to the changes in cell concentration near the edgeof the measuring spot, but later, chemotactic effects maycontribute (18). However, the initial rate of change in currentis linearly dependent on light intensity until light saturation isapproached and may therefore be used to measure actionspectra (Fig. 4).Complete spectra were routinely obtained by first record-

ing the response of the cells near an expected maximum inthe spectrum at relatively high actinic light intensity. Thelight intensity was then decreased by the insertion of neutraldensity filters until the response was in the linear range,which typically required at least a 1.0-optical-density filter.Then the wavelength was changed, in 10-nm steps usually,and in 5-nm steps in critical regions. For spectral regions inwhich less actinic light intensity was available, neutraldensity filters were removed and the responses were re-corded for at least one wavelength with and without the filterto ensure that linearity was maintained. For each measure-ment, the measuring spot was moved to a new area of thechamber with the specimen stage. The experiments showedthat neither the position in the chamber nor the depth atwhich the objective lens was focused was critical.The noise in the measurements was mainly determined by

two factors: the random movements of cells into and out ofthe measuring spot and the active movements of cellsswimming within the measuring spot. This latter part of thenoise was also detected in shallow chambers, when the cellswith their flagella stuck to the surfaces of the chamber butwere still rotating. As expected, as cell number increased,the noise increased to a maximum value and then decreased.With dead cells, which collected mainly at the bottom of thechamber, no significant noise was present at the signalamplification used.A complication arose because the response of the cells in

the chamber decreased slowly over several hours, for still-unknown reasons. To correct for this effect, the measure-ment at the first recorded wavelength was used as a standard

J. BACTERIOL.

on May 22, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

TECHNIQUE FOR OBTAINING PHOTOTAXIS ACTION SPECTRA

0 5 10t (min)

15 20

FIG. 3. Phototactic responses of H. halobium. (A) Depletion of cells in the measuring spot after 10 min of exposure to blue light (SchottBG3 filter) through the incident light illuminator and to orange background light (Schott OG 515 filter) through the dark-field condenser.Standard 100-W microscope lamps were used. (B) Accumulation of cells when the measuring spot was illuminated with orange light (SchottOG 515 filter). The blue dark-field illumination was applied only during the photographic exposure to show the cell population outside the spot.

(C) Typical recording of an attractant response to 590-nm light. , Light on; I light off; IPM, photomultiplier current.

and repeated after every five measurements. The decrease inresponse was usually found to be linear and amounted to as

much as a factor of two in 3 to 4 h. The responses at allwavelengths were normalized to the nearest standard wave-

length. At much-longer times, the response began to de-crease suddenly and dropped to near zero within 10 to 15min. Typical recordings show unexpected undershoots andovershoots of the traces during the first 20 s after the actiniclight had been turned on or off (Fig. 5). Possible explanationsare considered in Discussion. We measured as the initial

rates the slope of the curves after the under- or overshootswhen the traces crossed the dark level.Data for a complete action spectrum from 350 to 680 nm

(Fig. 6A) was obtained in approximately 2.5 h from the timethe cell suspension was sealed in the specimen chamber. Thestrain used, Flx3, contains only minor amounts of sR-II. Asexpected from previous observations on single cells (17), a

pronounced repellent effect of short-wavelength light was

obtained only with a long-wavelength background illumina-tion. The action spectrum corresponded well to the absorp-

C

t590 nm I

2793VOL. 170, 1988

on May 22, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

2794 STOECKENIUS ET AL.

4

3

~12 *3-

4221 /.1~~~~~~~~~~~~~~~~~~_1

0.O1s 4 8

0 1 2 3 4 5 6 7 8 9 10

'PM

FIG. 4. Light intensity dependence of the attractant response at590 nm. The arrows indicate the linear range used to obtain thespectra. The inset shows the same data plotted on a semilogarithmicscale. IPM, Photomultiplier current.

tion spectra of the phototaxis receptor pigment sR-I587 andits short-wavelength photoproduct S-I373 (24). The spectrumin Fig. 7 was obtained with a strain that has lost the responseto sR-I587 but retained the response to sR-Il480. For moststrains, contributions from both receptors are seen. Fornoise reduction, several action spectra were averaged. Thevariations observed for individual spectra are shown in Fig.6B for the attractant response.

DISCUSSIONIn general the action spectra obtained with the method

described here are very similar to the spectra obtained by

FIG. 5. Nonsaturating attractant (A) and repellent (B) responsesat different wavelengths of attractant light with the xenon lamp andmonochromator. Arrows mark the illumination periods (as de-scribed in the legend to Fig. 4). Note the marked overshoots. Thesteep changes in intensity following each transition were caused bymoving the specimen to a new position not previously illuminated.IPM, Photomultiplier current.

0.

t- A

-2 o.01

-3- i Iq'1 0 , 1

'\ /

-2- o°

-4 \ I'Q05p~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~

400 500Wavelength nm

e60

Wavelength nm

FIG. 6. (A) Complete action spectrum for H. halobium. Flx3cells without (0) or with (0) background light (>520 nm). 9

Absorption spectrum of sR-I587; ---, absorption spectrum ofphotointermediate S-I373 (spectra were taken from reference 24).The spectra were normalized at 590 and 370 nm, respectively. Fordetails, see text. (B) Action spectrum for the attractant responseobtained by summing eight single spectra. A retinal-negative mutantof H. halobium (Flx3R) grown in the presence of retinal was used.The bars mark the maximum and minimum values observed in thesingle spectra. -, Absorption spectrum of sR-I587.

others from single-cell observations; discrepancies in theinterpretations of the spectra have recently been discussed[19; W. Stoeckenius, E. Hildebrand, and A. Schimz, Letter,Trends Biochem. Sci. 11:402-403, 1986).

In principle, it should be possible to derive the lightresponse of the population quantitatively from the responses

0uuWavelength nm

FIG. 7. Action spectrum for the repellent response of H. halo-bium Flx3 KMlDl, a strain which shows only the response ofsR-I1480. The absorption spectrum of the receptor pigment (-) wastaken from reference 24. The spectrum was normalized at 480 nm.

J. BACTERIOL.

on May 22, 2018 by guest

http://jb.asm.org/

Dow

nloaded from

TECHNIQUE FOR OBTAINING PHOTOTAXIS ACTION SPECTRA

of single cells, which are due to modulations of their spon-taneous reversal frequencies by changes in the light intensity(18). The derivation is not possible, so far, because theconditions near the edge of the measuring spot, e.g., the lightgradient, are not well defined enough and the conditionsunder which single-cell reactions have been observed arequite different. Qualitatively, the observed response wasroughly what one would expect from single-cell observa-tions, except for the over- and undershoots of the responseswhen the actinic light was turned on and off. Of theseresponses, the paradoxical off reactions were much morepronounced and may be explained, at least in part, by therecently discovered fact that cells, after a light-inducedreversal, are refractory to a reversal-inducing stimulus andrecover their reactivity only slowly after several seconds(10-13). Cells swimming inside the actinic spot near its edgein an outward direction will reverse when an attractant lightis turned off. Cells swimming inward have just experienced areversal at the edge and will not react to the decrease inattractant light intensity. Thus, when the light is turned off,all cells will swim inward, and since the effective light spotdue to light scattering is somewhat larger than the image ofthe condenser aperture, a transient increase in light scatter-ing intensity should be observed. Indeed, a small increase ofthe photometer aperture so that it included a narrow ringoutside the condenser aperture image largely abolished theovershoot. An analogous argument can be used to explainthe overshoot in the depletion effect when a repellent light isturned off. The much less pronounced paradoxical responsesseen after the actinic light had been switched on may simplybe due to a change in the interval between reversals, whichwould be expected to cause a transient small decrease in cellnumber in the measuring spot for attractant light and anincrease for repellent light.The response pattern may be further complicated by the

summation of simultaneous stimuli and the inverse re-sponses seen in single cells under certain conditions (10, 11).Since we used the initial rate of the scattering changes, oursignal was determined only by cell concentration changes inthe periphery of our measuring spot. The central part onlyraises the light level and contributes noise. An obvious wayto increase the signal-to-noise ratio would be to use one ormore concentric rings instead of a solid circle of light.Preliminary experiments confirmed this view. However,implementations of this approach require the manufacture ofspecial apertures and modifications of the condenser andphotometer; the optimal arrangement still remains to bedetermined.

ACKNOWLEDGMENTS

Most of this work was carried out while Walther Stoeckenius wason sabbatical leave at the Max Planck Institute, Dortmund, andsupported by a Humboldt Prize. The work in San Francisco wassupported by Public Health Service grant GM-27057 from theNational Institutes of Health.We are grateful to Eberhard Waechter of E. Leitz, Cologne,

Federal Republic of Germany, for the loan of equipment andtechnical help and to Gisela Latzke for maintaining the cell strainsand preparing our samples.

LITERATURE CIT:ED1. Baryshev, V. A., A. N. Glagolev, and V. P. Skulachev. 1981.

Sensing of ACIH+ in phototaxis of Halobacterium halobium.Nature (London) 292:338-340.

2. Dencher, N. A. 1983. The five retinal-protein pigments ofhalobacteria: bacteriorhodopsin, halorhodopsin, P 565, P 370,and slow-cycling rhodopsin. Photochem. Photobiol. 38:753-767.

3. Dencher, N. A., and E. Hildebrand. 1979. Sensory transductionin Halobacterium halobium: retinal protein pigment controlsUV-induced behavioral response. Z. Naturforsch. Sect. C34:841-847.

4. Diehn, B. 1969. Action spectra of the phototactic responses ineuglena. Biochim. Biophys. Acta 177:136-143.

5. Dundas, I. E. 1977. Physiology of Halobacteriaceae. Adv.Microbiol. Physiol. 15:85-120.

6. Engelmann, T. W. 1883. Bacterium photometricum. Ein Beitragzur vergleichenden Physiologie des Licht- und Farbensinnes.Pfluegers Arch. Gesamte Physiol. Menschen Tiere 30:95-124.

7. Hader, D. P. 1979. Control of locomotion, p. 268-309. In W.Haupt and E. Feinleib (ed.), Encyclopedia of plant physiology.New series: physiology of movements. Springer-Verlag KG,Berlin.

8. Hader, D. P., and M. Lebert. 1985. Real time computer-controlled tracking of motile microorganisms. Photochem. Pho-tobiol. 42:509-514.

9. Hartmann, K. M. 1983. Action spectroscopy, p. 115-144. In W.Hoppe, W. Lohmann, H. Markl, and H. Ziegler (ed.), Biophys-ics, 2nd ed. Springer-Verlag KG, Berlin.

10. Hildebrand, E., and A. Schimz. 1986. Integration of photosen-sory signals in Halobacterium halobium. J. Bacteriol.167:305-311.

11. Hildebrandi E., and A. Schimz. 1987. Role of the responseoscillator in inverse responses of Halobacterium halobium toweak light stimuli. J. Bacteriol. 169i254-259.

12. Marwan, W., and D. Oesterhelt. 1987. Signal formation in thehalobacterial photophobic response mediated by a fourth retinalprotein (P480). J. Mol. Biol. 195:333-342.

13. McCain, D. A., L. A. Amici, and J. L. Spudich. 1987. Kineticallyresolved states of the Halobacterium halobium flagellar motorswitch and modulation of the switch by sensory rhodopsin I. J.Bacteriol. 169:4750-4758.

14. Nultsch, W., and M. Hader. 1978. Photoakkumulation bei Ha-lobacterium halobium. Ber. Dtsch. Bot. Ges. 91:441-453.

15. Spudich, E. N., and J. L. Spudich. 1982. Control of transmem-brane ion fluxes to select halorhodopsin-deficient and otherenergy-transduction mutants of Halobacterium halobium. Proc.Natl. Acad. Sci. USA 79:4308-4312.

16. Spudich, J. L., and R. A. Bogomolni. 1983. Spectroscopicdiscrimination of the three rhodopsinlike pigments in Halobac-terium halobium membranes. Biophys. J. 43:243-246.

17. Spudich, J. L., and R. A. Bogomolni. 1984. The mechanism ofcolour discrimination by a bacterial sensory rhodopsin. Nature(London) 312:509-513.

18. Spudich, J. L., and W. Stoeckenius. 1979. Photosensory andchemosensory behaviour of Halobacterium halobium. Photo-biochem. Photobiophys. 1:43-53.

19. Stoeckenius, W. 1985. The rhodopsin-like pigments of halobac-teria: light energy and signal transducers in an archaebacterium.Trends Biochem. Sci. 10:483-486.

20. Sundberg, S. A., R. A. Bogonlolni, and J. L. Spudich. 1985.Selection and properties of phototaxis-deficient mutants ofHalobacterium halobium. J. Bacteriol. 164:282-287.

21. Takahashi, T., and Y. Kobatake. 1982. Computer-linked auto-mated method for measurement of the reversal frequency inphototaxis of Halobacterium halobium. Cell Struct. Funct.7:183-192.

22. Takahashi, T., T. Tomioka, N. Kamo, and Y. Kobatake. 1985. Aphotosystem other than PS370 also mediates the negative pho-totaxis of Halobacterium halobium. FEMS Microbiol. Lett.28:161-164.

23. Takahashi, T., M. Watanabe, N. Kamo, and Y. Kobatake. 1985.Negative phototaxis from blue light and the role of thirdrhodopsinlike pigment in Halobacterium cutirubrum. Biophys.J. 48:235-240.

24. Wolff, K. E., R. A. Bogomolni, P. Scherrer, B. Hess, and W.Stoeckenius. 1986. Color discrimination in halobacteria: spectro-scopic characterization of a second sensory receptor coveringthe blue-green region of the spectrum. Proc. NatI. Acad. Sci.USA 83:7272-7276.

VOL. 170, 1988 2795

on May 22, 2018 by guest

http://jb.asm.org/

Dow

nloaded from