Fireflies at One Hundred Plus: A New Look at Flash Control1

203

INTEGR. COMP. BIOL., 44:203–212 (2004)

Fireflies at One Hundred Plus: A New Look at Flash Control1

HELEN GHIRADELLA2 AND JOHN T. SCHMIDT

Department of Biological Sciences, The University at Albany, Albany, New York 12222

SYNOPSIS. The mysterious process by which fireflies can control their flashing has inspired over a centuryof careful observation but has remained elusive. Many studies have implicated oxygen as the controllingelement in the photochemical reaction, and the discovery of nitric oxide synthetase (NOS) in the lantern hassuggested that nitric oxide (NO) may control oxygen access to the light-emitting photocytes, thereby trigger-ing the flash. However, there are several drawbacks to oxygen as a controlling agent, and in view of theprominence of peroxisomes in lantern morphology and biochemistry, we suggest that it is hydrogen peroxidethat triggers the flash, and we present a model by which this may take place.

INTRODUCTION

Any discussion of the biology of fireflies inevitablyincludes the mystery of their light production. Overthe years a model of the biochemistry of firefly lightproduction has evolved (see Wilson and Hastings,1998, for review), a model that has identified oxygenas the major control factor. In simple terms, the modelstates that firefly luciferin (the substrate) complexeswith firefly luciferase (the enzyme) and ATP in thepresence of Mg21 to form an ‘‘active intermediate’’that needs only oxygen to complete the photochemicalreaction. With the addition of the oxygen, the activeintermediate forms a cyclic peroxide that decomposesand in the process emits light. When biochemists mixthese chemicals together, they typically produce asteady glow unless oxygen is added last, in which caseone gets a flash that resembles that of the in vivo flash(DeLuca and McElroy, 1974). As a group, fireflies aremore versatile: most can modulate their light to agreater or lesser extent (see Buck [1948] for a thor-ough overview of firefly anatomy and physiology),and some can actually flash, turning the lantern sharplyon and off (Fig. 1). This singular ability has provokedover a century of careful morphological and physio-logical search for the elusive control switch, the eventthat pushes the photochemical reaction to completion.This paper reviews current models of flash control andadds to them a new twist: we review the biochemistryof the light reaction and ask if, rather than oxygen, itmight be hydrogen peroxide that actually triggers theflash.

Various lines of evidence (reviewed in Ghiradella,1998) have certainly supported the idea that flash con-trol is essentially oxygen control. Stop-flow experi-ments (DeLuca and McElroy, 1974) demonstrated thatthe addition of oxygen to the preformed intermediateis sufficient to produce light with a speed (on the orderof 60 msec) close to the rise time of lantern light.Within the light-emitting cells, or photocytes, the lu-ciferin and luciferase are contained within peroxi-

1 From the Symposium Flash Communication: Fireflies at Fiftypresented at the Annual Meeting of the Society for Integrative andComparative Biology, 4–8 January 2003, at Toronto, Canada.

2 E-mail: [email protected]

somes (present in huge numbers), and not surprisingly,it is the peroxisome fraction of the cell that lights up.Peroxisomes are subcellular organelles noted for theirenzymatic oxidation reactions, which are coupled tothe formation of hydrogen peroxide from oxygen.They also contain catalase, which reduces hydrogenperoxide to water (either with hydrogens from coupledreactions, or without other substrates by simultaneous-ly oxidizing and reducing the peroxide to oxygen andwater). The lantern nerves, which appear to be modi-fied spiracular nerves, do not synapse directly on thephotocytes, but rather in the tracheal system, which ininsects delivers air directly to the tissues. In short, thecircumstantial evidence has pointed to oxygen as a ma-jor player in flash control. However, for a long timewe have had no clear picture as to how molecular ox-ygen could be controlled or gated by a biological sys-tem.

Two recent studies have provided us with new pos-sibilities. To understand these we must first reviewbriefly the unique morphology of a typical flashinglantern, taking that of an adult Photuris as our model.We will concentrate on those structures known or be-lieved to be involved in the flash and its control; read-ers wishing more comprehensive views of lantern mor-phology are urged to consult Buck (1948) for generallantern anatomy, Smith (1963) for an excellent generaldescription of lantern ultrastructure in Photuris, andGhiradella (1998) for a more recent review of lanternmorphology and biochemistry and a detailed look atthe tracheal system and its environs.

MORPHOLOGY OF THE PHOTURIS LANTERN

Each lantern is a flat slab of tissue that consists ofa dorsal and a ventral layer. The dorsal layer consistsof large cells packed with granules whose nature is asyet undetermined (but which have been rumored byvarious authors to be urate, possibly because urate isa common feature of fat body from which the lanternappears to be derived). The ventral, or photogenic lay-er is the source of the light and is our concern here.Figure 2 presents a view in frontal section of the ven-tral layer of an adult Photuris lantern. The photocytesare arranged in ‘‘rosettes’’ around channels or ‘‘cyl-inders,’’ each of which contains branches of the tra-

Dow

nloaded from https://academ

ic.oup.com/icb/article/44/3/203/600893 by guest on 09 M

arch 2022

204 H. GHIRADELLA AND J. T. SCHMIDT

FIG. 1. Photuris versicolor, characteristic flash pattern (upper trace)and associated volleys in the cord (lower trace). Flash rise time(between vertical bars) is on the order of 70–75 msec; calibrationmark at bottom is 0.1 sec. Courtesy of A. D. Carlson.

FIG. 2. Photuris sp., light micrograph of a frontal section of thelantern, showing several cylinders and their accompanying photo-cytes (compare Fig. 3). The main tracheae and their epithelial cellsare the signet ring profiles in the centers of the cylinders; some ofthe finer vessels show as tiny open profiles at the cylinder borders.The prominent nuclei lining the borders are those of tracheolar andoccasionally tracheal end cells. An arrow indicates one of the lanternnerves. Individual photocytes stretch from cylinder to cylinder: theirdifferentiated zones appear as faint streaks between them and as lightbands at the borders of the cylinders. The coarse granules in thephotocyte interiors are the peroxisomes. Bar 5 10 mm.

cheal system and of the lantern nerve. Each photocyteabuts on at least two cylinders, so that it receives tra-cheal and nervous inputs at both ends. Those regionsof the photocyte that front on the cylinders or on thetracheal extensions (see below) are distinctly differentfrom the more interior parts of the cell and are calledthe ‘‘photocyte differentiated zones’’ (Fig. 3). Theycontain virtually all of the cell’s mitochondria, as wellas so-called ‘‘differentiated zone granules’’ (DZ gran-ules), an as yet uncharacterized organelle that is alsofound deeper in the cell. The most striking feature ofthe cell’s interior is the high concentration of peroxi-somes, which virtually fill the cell to the exclusion ofanything else other than the nucleus. As mentionedabove, these organelles contain the enzyme and sub-strate for the light reaction, and it is this fraction ofthe cell that actually emits light.

The lantern tracheal system is also highly special-ized. All insect tracheal systems are essentially exten-sions of the outer exoskeleton, which invaginates intothe interior, carrying with it its epithelial layer, whichlatter encases the tubes thus formed. Typical tracheaebranch repeatedly until each branch is so fine that it isensheathed in a single epithelial cell. This ‘‘terminal’’trachea then gives rise to a spray of delicate trache-oles, all encased in the finger-like branches of a singletracheolar cell. The tracheoles are the capillaries ofthe system: they are permeable and are usually delicateand collapsible, despite the presence of taenidia, cu-ticular threads that usually wrap around respiratorytubes and presumably provide mechanical support. Itis the tracheoles (and the attendant branches of theirtracheolar cells) that front on and penetrate betweenthe tissue cells to ensure that these have close accessto gas exchange.

The firefly lantern tracheation is highly modified(Figs. 3, 4). It is perhaps the most extensive of anyknown tissue. The last of the tracheal epithelial cells(just upstream of the tracheolar cell) is now called thetracheal end cell (TEC); it is greatly enlarged and itsluminal surface thrown into deep crypts lined with mi-tochondria. This is classic ion pump morphology, char-acteristic of tissues that move water by shifting suchions as sodium (see Smith, 1968, for review), and wewill proceed on the assumption that this is its function,

although this hypothesis has not been tested or con-firmed. The TEC wraps around the tracheolar cell,which also has a hint of ion pump morphology, andthe nerve ending is located between the two, embrac-ing and (in our images at least) synapsing on the tra-cheolar cell. The transmitter is octopamine, commonin invertebrate systems (Oertel and Case, 1976; Na-thanson, 1979).

The tracheoles, with their attendant cell branches,penetrate far between the photocytes and end blindlynear the ends of their opposite numbers from theneighboring cylinder. As mentioned above, where thetracheal system abuts on or runs past a photocyte, itis accompanied by differentiated zone in the latter. Par-adoxically, the intercellular gap between tracheal sys-tem and photocyte is large and seems to be reinforcedby fibrous material, perhaps elements of basementmembrane.

The cuticular elements of the lantern tracheal systemare also highly specialized (Fig. 4). The tracheae aremore or less standard, but between each of their ter-minal branches and its tracheoles lies an unusually fea-tureless segment of tracheal tubing, the tracheal twig,which is lacking in taenidia and appears particularly

Dow



arch 2022

205FIREFLY FLASH CONTROL

FIG. 3. Diagram of an end organ and its environs (not all elements drawn to scale). The incoming trachea becomes a fine tracheal twig(compare Fig. 4), surrounded by the tracheal end cell with its putative pump morphology. The tracheal twig puts forth a spray of tracheolesthat lie within the tracheolar cell. Tracheoles and tracheolar cell branches extend deep between the photocytes. The nerve synapses betweenthe tracheal end and tracheolar cells, with synaptic profiles onto the latter. Each photocyte is differentiated into mitochondria-filled ‘‘differ-entiated zones,’’ which line the tracheal system, and the peroxisome-filled interior. An as-yet uncharacterized granule is also found in photocytes,particularly in the differentiated zones. Modified from Ghiradella, 2004.

flimsy and collapsible in whole mount. The tracheoles,in contrast, are uniquely elaborate, stiff and reinforced.

The lantern tracheal end cell, tracheolar cell andnerve ending, together with their enclosed plumbing,are collectively referred to as the ‘‘end organ.’’ Itseems that these specializations, together with thecompartmentalized photocyte, are necessary for a lan-tern to flash, since only flashing lanterns have themall. We find a ‘‘control’’ lantern in the Photuris larva,which produces a glow that slowly rises and falls. (Fora review of larval lantern ultrastructure and physiolo-gy, see Oertel et al., 1975.) The larval photocyte alsohas peroxisomes, mitochondria (and DZ granules), butthese are not segregated from one another in differentcellular compartments, nor is the tracheal system par-ticularly specialized. The nerve synapses directly onthe photocyte, rather than on the tracheolar cell. Thebiochemistry of the light reaction seems similar in lar-val and adult organs (Oertel et al., 1975), but theseare not derived from one another and in fact coexistbriefly at the time of eclosion, so it is not known towhat degree they are homologous or how far conclu-sions drawn from one can be generalized to the other.

OXYGEN AS A POSSIBLE AGENT OF FLASH CONTROL

The general acceptance of oxygen as the controllingfactor for the adult flash inspired suggestions (re-

viewed in Ghiradella, 1998, 2004, and Greenfield,2001) as to possible mechanisms. All had features incommon: they posited that the peroxisomes arecharged with active intermediate that needs only a puffof oxygen to complete the flash reaction. The questionthen became what keeps the oxygen from entering thereaction prematurely, i.e., how does the interior of thephotocyte stay anaerobic enough to avoid setting offthe light reaction? On the basis of photocyte mor-phology, Ghiradella (1998, and in papers referencedtherein) and before her, Dahlgren (1917, in a remark-ably prescient paper) suggested that because of theirdeployment at the tracheal system/photocyte interface,the photocyte mitochondria might serve as ‘‘gatekeep-ers,’’ absorbing and using any incoming oxygen tokeep it from the peroxisomes until the nerve fires andthe octopamine sets in motion a cascade of events thatdrops the barrier and allows the oxygen to reach theperoxisomes and complete the photochemical reaction.

Trimmer et al. (2001) provided a possible mecha-nism for this action when they realized that the gas-eous transmitter nitric oxide (NO) might be the neededlink between events at the synapse and the rise of theflash. In Figure 4 (center) the shaded areas (the TEC,the tracheolar cell and its extensions, and differentiatedzones of the photocytes) represent sites of nitric oxide

Dow



arch 2022


FIG. 4. Center diagram, as Figure 3, showing the location of the nitric oxide synthetase (grey). This figure also shows details of the trachealsystem: A. Whole mount (negatively stained) of parts of a trachea (T), tracheal twig (tt), and its tracheoles (t). The tracheal twig shows thecloud of fibrils that often accompanies this structure, even in washed preparations. The elaborate structure of the tracheoles (see also D) is insharp contrast to the featureless tracheal twig. The asterisk (*) shows for comparison a typically simple and unadorned non-lantern tracheole,which somehow made its way into the preparation. Bar 5 0.5 mm.

B. End organ region, transmission electron micrograph. The trachea (T) gives rise to the tracheal twig (tt), surrounded by its calyx of fibrils,and thence to the tracheoles (t—no longer in the plane of the section). The crypts of the tracheal end cell (TEC) appear to converge on thefibrillar area of the twig. Bar 5 1 mm.

C. Scanning electron micrograph of a critical point dried preparation of several tracheae (T), tracheal twigs (tt) and tracheoles (t). The spiraltaenidia that surround tracheae and tracheoles are clearly missing from the twigs. The tracheoles are characteristically stiff, one clear special-ization of the lantern. Bar 5 2 mm.

D. Whole mount, similar to A. Here a tracheal twig (outlined by white arrows) happened to fall across two tracheoles; its flimsiness isapparent, as is the ornate specialization of the tracheoles. Bar 5 0.5 mm.

Dow



arch 2022


synthetase (NOS). In their model, Trimmer et al. sug-gested that the NOS is activated and produces NO inresponse to calcium transients raised by the action ofoctopamine at the cell surface. NO reversibly inacti-vates mitochondria by complexing mitochondrial cy-tochrome c oxidase, without which the mitochondriacannot bind oxygen. With the mitochondria ‘‘off,’’ ox-ygen can diffuse through the photocyte differentiatedzones to the peroxisomes. The authors extend the‘‘gatekeeper’’ hypothesis to the TEC and suggest thatits prominent mitochondria might serve as a first (out-lying) line of defense against oxygen invasion until theNO in the TEC shuts down its mitochondria and al-lows additional oxygen access to the photocyte.

Built into this model are welcome additions to pre-viously suggested mechanisms for flash termination. Invitro studies by McElroy and Hastings (1956) sug-gested that the flash might be self-limiting because theformation of active intermediate is much slower thanthe flash reaction itself. Trimmer et al. point out thatNO degrades and diffuses away quickly, it is inhibitedby oxygen (which will build up when the mitochondriacease absorbing it), and it is inhibited by light itself(Aprille et al., 2002, 2004). All these factors can beexpected to contribute to flash termination.

The NO model does not, however, address the spe-cializations of the tracheal system, notably the ionpump morphology of the TEC and tracheolar cell, thereinforcement of the tracheoles and the existence ofthe flimsy tracheal twig. A different group of investi-gators (Timmins et al., 2001) has focussed on theseelements and revived an old osmotic model of oxygencontrol (Maloeuf, 1938) that is based on presumedshifts by the ion pump of tissue fluid in and out of thetracheoles, a process that mimics one that occurs gen-erally in insect tissues as a result of metabolic activity(Wigglesworth, 1983).

In resting insect tissues, tracheoles are typicallyfilled with liquid drawn by capillary action from thesurrounding tissue. Metabolic activity raises tissue os-motic pressure, which in turn draws the liquid backfrom the tracheoles, filling them with air. Since oxygendiffuses more rapidly by a factor of 104 in air than inwater, this allows much faster oxygen delivery to theneedy active tissues. Timmins et al. studied the kinet-ics of oxygen diffusion in tracheoles under differentconditions and concluded that the singular morpholo-gies of the TEC and tracheolar cell suggest that theymight respond to the octopamine by actively pumpingtissue fluid out of the tracheoles, thereby clearing themfor fast oxygen passage. The burst of incoming oxygenwould overwhelm mitochondrial ability to absorb itand would reach the peroxisomes, initiating the flash.Contrary to the conventional wisdom that pumping/osmotic mechanisms are slow, these authors assert thatfor such short distances the process could be fastenough to work within the required time frame. Theyalso echo an older idea (Ghiradella, 1977) that rein-forcement of tracheoles and intercellular gaps would

keep the tubes and their attendant cells stable undershifting osmotic pressures.

The osmotic control model of Timmins et al. doesnot contradict or preclude an NO effect on the pho-tocyte mitochondria. In a combined model, oxygendiffusing rapidly through air-filled tracheoles would ar-rive at a photocyte ‘‘opened’’ by NO shut-down of themitochondrial gatekeepers and would quickly and ef-fectively reach the peroxisomes. More troubling is thelack of mechanism connecting the octopamine with theosmotic activity at the TEC. If NO shuts down pho-tocyte mitochondria, it should also shut down TECmitochondria, presumably turning off the pump. Theremay be enough stored ATP to run the pump tempo-rarily (J. R. Aprille, personal communication), but weare then left with the question of what difference theoctopamine makes in TEC function, i.e., why this cellhas NOS if the pump is on regardless of the presenceor absence of NO. It is more likely that TEC pumpingis important when the lantern is off or that it may servea function not directly related to control of individualflashes.

The crypts of the TEC are oriented towards the tra-cheal twig. On the basis of this morphology Ghiradella(2003) revived an old suggestion that in the trachealtwig the system might have a mechanical valve whichcould open or close in response to changes in the TEC.(The reinforcement of the intercellular gap and the tra-cheoles would then be against mechanical as well asosmotic stress.) In whole mounts (Fig. 4) the twig isalways collapsed shut, but in fixed and sectioned ma-terial, it always appears open. This latter fact is notnecessarily informative in itself, since fixed material issubject to shrinkage, but the twig is typically surround-ed by a calyx of fibrils which appear to come from thecrypts of the luminal surface of the TEC and whichmay bind TEC and twig together. Osmotic changes inthe region should swell or shrink the TEC, and this inturn might be a mechanism for opening or shutting thetracheal twig (and facilitating or delaying the passageof oxygen). But we still need to know why and underwhat circumstances.

Some additional special features of the lantern de-serve mention. It has an extremely potent octopamine-sensitive adenyl cyclase (Nathanson, 1979), which ap-pears to be localized in the tracheal system and whichproduces the highest cAMP levels recorded for anyinsect (Zeng et al., 1996). Case and Strause (1978)point out that although the adult lantern has relativelysparse innervation, action potentials are easily record-ed from the surface, suggesting that another compo-nent—perhaps the tracheolar cells—may be capable ofgenerating spikes. Certainly these cells, with their longbranches, their close associations with the photocytedifferentiated zones and their ability to make NO, arewell situated to carry the message, whatever it may be,deep into the photocyte interiors. But all these spe-cializations must be incorporated into any model offlash control.

Dow



arch 2022


PROBLEMS WITH THE OXYGEN CONTROL THEORY

If external oxygen is a simple trigger, a cut or breakin the lantern cuticle or surface should light up thewhole lantern as oxygen rushes in through the break.This does not happen: unexposed parts of a cut orcrushed lantern do not glow (Dahlgren, 1917), and acut part glows only locally and quickly extinguishes(H. Gh., unpublished data). Other studies (Hastingsand Buck, 1956; Carlson, 1965) suggest strongly thatthe oxygen supply within the lantern does not limit orcontrol the glow. Furthermore, the lantern is not in anyway isolated from the rest of the body. It floats incirculating hemolymph that is presumably oxygenatedfrom non-lantern tracheal sources. It seems clear thatthe photocyte has substantial local control of its lu-minescence and is not dependent on external oxygenin any simple way. Let us take a closer look at thiscell, and in particular at the peroxisomes that fill it sorobustly.

Implicit in the whole oxygen control model has beenour assumption that the photocyte must work hard tokeep its interior anaerobic between flashes. This hasseemed to be a unique and exotic qualification, but asLane (2002) points out, all eukaryotic cells are func-tionally anaerobic. Indeed, mitochondria may haveoriginally evolved to ‘‘compartmentalize’’ respiration,keeping the oxygen levels in the cytosol just highenough to support oxygen demands while minimizingthe exposure to potentially dangerous reactive oxygenspecies (ROS) (Abele, 2002). In other words, anyhealthy cell can presumably keep itself safe from ox-ygen overload and the cytotoxic species (superoxide,peroxynitrite, hydrogen peroxide and [more recentlydescribed] ozone, dihydrogen trioxide, and the hydro-trioxy radical) that are likely to result from it. Giventhe great ability of biological systems to build on andmodify common systems for exotic purposes, if oxy-gen levels needed to be especially manipulated, onewould expect some sort of mitochondrial elaborationto provide the needed control.

But it is the peroxisomes in the lantern that lightup, and while they come in a variety of forms, per-oxisomes have a specific talent—they can manipulateperoxides. They typically form hydrogen peroxide andsuperoxides, which they generate through various per-oxidative reactions, and they then use catalase, the oneenzyme common to them all (de Duve, 1969) to breakthe hydrogen peroxide down to oxygen and water. Fur-thermore, they do this safely, i.e., without productionof free radicals (Lane, 2002). Theoretically, photocyteperoxisomes could store oxygen in the form of hydro-gen peroxide, liberating it as needed through the actionof catalase. But NO inhibits catalase (Brown, 1995)and therefore precludes this release of oxygen, whichinhibition might lower oxygen levels somewhat, evenas oxygen coming in from the tracheoles would raisethem. In other words, in the active lantern we wouldhave forces working in opposing directions, whichhardly seems an effective way of initiating an explo-

sively fast chemical reaction (the rise of the flash inthe insect can be even more rapid than that achievedwith a step pulse [something a biological system can-not produce] of oxygen in a stop-flow apparatus).

PEROXIDE AS THE AGENT CONTROLLING THE FLASH

If we assume that in the intact lantern the activatingspecies is not oxygen but hydrogen peroxide, thingslook very different. The various peroxisome oxidases(urate-, D,L amino acid-, glycolate-, and L-hydroxyacid oxidases) would use the burst of external oxygenarriving through the ‘‘open’’ mitochondria to make hy-drogen peroxide, and the shutdown of catalase wouldallow this compound to build up explosively, com-pleting the photochemical reaction and raising theflash. We need now to revisit the chemistry of the lan-tern light reaction and ask if hydrogen peroxide in factcan serve as this trigger.

We propose the following speculative scenario (Fig.5): The release of octopamine at the synapse both de-polararizes the tracheolar cell to fire an action potentialand activates the octopamine-sensitive adenyl cyclasein the end organ complex to form cAMP. The actionpotential depolarizes the tracheolar cell and all itsbranches, deep between the photocytes, opening volt-age dependent calcium channels (VDCC) in the cellmembrane. The cAMP acts on protein kinase A tophosphorylate calcium channels and augment the cal-cium influx (Nishiyama et al., 2003). The resulting risein intracellular calcium activates the NOS to produceNO throughout the whole tracheolar cell. The NO dif-fuses rapidly into the photocyte, where it shuts downthe mitochondria, opening the photocyte interior to in-coming oxygen. But it also reaches the peroxisomeswhere it shuts down the catalase. This is possible be-cause the Ki for catalase inhibition (180 nM) is evenlower than that for cytochrome c oxidase inhibition(270 nM: Brown, 1995). Hydrogen peroxide levels risesharply, abetted both by peroxide formation from theincoming oxygen and by the inhibition of catalase. Wepropose that it is the rise in the highly reactive per-oxide that completes the photochemical reaction. In-deed, one might expect such a fast burst of hydrogenperoxide production from the action of another com-mon peroxisomic enzyme, urate oxidase (assumingthat the lantern is rich, as it may be, in urate inheritedfrom its ancestral fat body).

In this model, the end organ and tracheolar cellbranches are in fact serving at least three purposes:they are delivering oxygen, they are carrying excita-tion rapidly along their lengths, and they are makingand broadcasting NO throughout the lantern. Thesefunctions can be related to the physiological differenc-es between species. Carlson (2003) has demonstratedeffectively that two species, Photuris versicolor, whichhas a compound flash, and P. lucicrescens, whose flashis a slowly rising crescendo, respond quite differentlyto similar nervous stimuli and to various high K1 andCa21 solutions. He concludes that there must be a fun-damental difference at the photocyte, possibly involv-

Dow



arch 2022


FIG. 5. The model mechanism for the control of the lantern flash. The nerve synapses onto the tracheolar cell, releasing octopamine (OA).Octopamine receptors (OAR) both depolarize tracheolar cells to fire spikes and also activate via the Gs protein (Gs) adenylate cyclase (AC),which makes cAMP from ATP. The activated cAMP-dependent kinase (PKA) can phosphorylate voltage dependent calcium channels to enhancethe ability of Ca21 to enter the cytoplasm (as well as other substrates that may enhance the excitability of this cell). The Ca21 binds tocalmodulin (Cam), activating nitric oxide synthetase (NOS) to release nitric oxide (NO), which can readily diffuse across cell membranes tocarry the message into the photocytes. (Photocytes may also have calcium channels and NOS for their own NO synthesis as well.) NO bindsto and inhibits cytochrome c oxidase, preventing its binding and use of oxygen; oxygen is then free to diffuse beyond the mitochondria to theperoxisomes, where it is converted into hydrogen peroxide. Hydrogen peroxide is normally quickly degraded by catalase, but NO also bindsand inhibits catalase, further contributing to the rise in peroxide levels. We hypothesize that it is the fast rise in hydrogen peroxide, ratherthan a slow rise in oxygen, that activates the luciferin-luciferase to emit light. As time goes by, NO is degraded by light and by action of thecatalase in the presence of peroxide and rising oxygen. These changes terminate the action of the NO, the transient rise in peroxide, and theflash.

ing voltage-dependent calcium channels (VDCC). Healso points out that Ca21 depolarizes mitochondria andcould presumably function directly in liberating oxy-gen from them as a result of octopamine action. Wesuggest that the tracheolar cell may be another candi-date for this level of control: if it is in fact capable ofspiking, its response to synaptic activity may differ inthe different species, with concommitant differences inthe flash pattern.

We can also further provide for flash shut-down. AsTrimmer et al. (2001) note, nitric oxide will diffuseout quickly, its binding to cytochrome c oxidase isinhibited by rising oxygen levels (caused by its inac-tivation of mitochondrial cytochrome c oxidase), andit is inhibited by light, all of which will render it in-

active. In addition, catalase can, in the presence ofhydrogen peroxide, degrade NO (Brown, 1995). AsNO drifts off and/or is degraded, oxygen levels in theperoxisomes will fall, and catalase activity in the per-oxisomes will rise, with the result that peroxide levelswill plummet, allowing the system to revert back toits quiet state where it will presumably make moreATP and break down any reaction products (oxyluci-ferin, AMP, CO2) that have accumulated.

All this depends, of course, on the effectiveness ofhydrogen peroxide as the trigger for the photochemicalreaction. The proposed chemistry of the peroxide-stim-ulated light reaction is shown in Figure 6. The lucif-erin-AMP intermediate (LH2-AMP) that was shown tobe present before the flash is activated by peroxide/

Dow



arch 2022


FIG. 6. Proposed chemistry of the light reaction, starting with the luciferin-AMP intermediate that forms in the absence of oxygen. Hydrogenperoxide attacks both the 69OH and the C4 carbon, removing two hydrogens. No free radicals are left at each site because the double bondsrearrange (arrows) to produce a neutral conjugated dehydro form of L-AMP (form 1). An alternative dehydro form (classical dehydro form 2)does not result in light emission. After peroxide attack on the double bond (step 2), the peroxy-bridge structure forms by attack on the carbonyl,driving off AMP (step 3). This structure then disintegrates, releasing CO2 and the ketone form of oxyluciferin, still in the rearranged form(C5O not conjugated with other double bonds). This spontaneously rearranges to give the original configuration (with the ketone now includedin the conjugated double bond system) and emits light.

hydroxyl radical attack on two hydrogens, first at theusual C4 carbon (site of postulated proton extractionin the oxygen model; H2O2 removing H• to make H2Oand •OH) and at the important 69OH (luciferin with amethyl group substituted here does not emit light:McCapra, 1970). Although this would nominally leavebehind radicals at both points, the formation of theseis averted by a concerted rearrangement of the doublebonds to give a neutral, conjugated L-AMP (dehydroform 1 of Fig. 6). Note that even the carboxyl doublebond is now included in the conjugated system. Oc-casional formation of the classical (unrearranged) de-hydro form 2 is unproductive for light formation(McElroy and DeLuca, 1978). The L-AMP interme-diate and a form without AMP apparently can disso-ciate from luciferase before they emit light (Min andSteghens, 1999; Brau et al., 2000). We envision furthersteps of peroxide attack at the double bond and for-mation of a peroxy-bridge before driving off AMP, andlater decomposition to the ketone, producing CO2 (asin the oxygen model). The difference is that this leaves

the new ketone double bond unconjugated with the restof the system. Simultaneous protonation at the 69O anddeprotonation of the nitrogen give rise to electronshifts associated with a rearrangement back into theoriginal configuration of conjugated double bonds, butwith the favorable addition of the ketone double bondto the conjugated system. (Alternatively this can bethought of as going to the enol structure [N5C-OH],which would rearrange after deprotonation of the hy-droxy group.) This model does not preclude the reac-tion of molecular oxygen facilitated by the luciferase,as obviously happens in the lab. But peroxide activa-tion, rather than oxygen activation, could explain theextremely rapid rise of the flash in the insect.

In presenting this admittedly speculative model wehave wished to launch a new direction in thinkingabout flash control in the firefly. We believe that wehave made a strong case for hydrogen peroxide as thekey agent in flash control, but we acknowledge thatthe system might not be that simple. Current researchis showing that oxygen can exist in more forms than

Dow



arch 2022


have previously been suspected (Hoffman, 2004) andthat biological systems can handle and control manyof them (Lerner and Eschenmoser, 2003; Wentworthet al., 2003). Besides singlet oxygen and hydrogenperoxide, already implicated in the firefly flash reac-tion, these ‘‘manageable species’’ include ozone, dih-ydrogen trioxide (H2O3), and the hydrotrioxy radical(HO3•); in antibody action, this last may produce ormay mimic the action of the notoriously toxic HO•radical. The field is advancing quickly, and as newknowledge accumulates, we shall almost certainlyhave to revisit the question of reactive oxygen speciesin firefly flash control.

Several additional questions remain. What of theTEC and its complicated morphology? As mentionedabove, it could serve as a mechanical valve, helpingto hold oxygen out of the lantern between flashes. Al-ternatively, it may be doing something similar on alonger time scale. A ‘‘daytime’’ firefly cannot flashwhen it is first awakened: there are seconds or evenminutes of disorganized lantern activity before the sys-tem can begin to produce coordinated flashes. This hasbeen generally interpreted as a need for some sort ofneural priming, but it may be that a physiological ad-justment needs to be made, such as (for example) theTEC turning on the pump and clearing the tracheolesof tissue fluid in preparation for the coming lanternactivity. The presence of NOS in the TEC may alsobe related in some as yet unknown fashion to the‘‘off’’ process in the flash.

We need to know more about the other enzymes inthe lantern peroxisomes. As a class peroxisomes havebeen found to have a huge variety of oxidases, manyof which have been selectively shed in various line-ages so that today’s organelles are more or less tailoredfor their specific environments and functions. As men-tioned above, urate oxidase may be of particular in-terest here: we need to determine if urate is actuallypresent in the lantern and being used to form the hy-drogen peroxide that we are proposing as the controlfor the light reaction.

We need more studies such as that of Carlson (2003)into control mechanisms for extended, compound andmultiple flashes (Fig. 1). We believe that we can safelyassume that the underlying biochemistry of the flashreaction is conserved among firefly species, but thedetails of flash patterning (especially in those specieswith complex flash repertoires) add a level of com-plexity that must ultimately be explained by any modelof flash control.

Finally, drawing on this model, can we say anythingabout the function of the larval lantern? We do notknow if there is NOS somewhere in the larval system.There certainly is no end organ with its specializedmorphology. The photocyte membrane is highly con-voluted and receives many nerve terminals whose lib-eration of octopamine leads to a slow depolarizationthat may involve cAMP; repeated stimuli facilitate thelight response (Oertel and Case, 1976). Unlike theadult, which can glow for several minutes in a nitrogen

atmosphere, the larva can do so for only a few seconds(Carlson, 1965), which has been interpreted to meanthat it has limited oxygen stores in the lantern. How-ever, if we think peroxide rather than oxygen, we cansuggest that lacking extensive tracheation to bring inoxygen for mitochondrial function at rest and the gen-eration of peroxide during the flash, the larval perox-isomes may have relatively low peroxidatic activityand so be unable to accumulate enough peroxide rap-idly enough to mount a flash.

We hope that our model will provoke furtherthought, especially in light of current research into thecomplexity and versatility of biological control of ox-ygen in its many forms and compounds. We can thinkof no more exciting or rewarding quest than that tounderstand how a biological system can evolve suchability to control the release of stored light.

ACKNOWLEDGMENTS

First and foremost, we are most grateful to JohnBuck, who first introduced one of us (H. Gh.) to thecomplexities of the firefly lantern and who, togetherwith his wife Elizabeth, has been a constant source ofhelp and support. We thank James Case, Al Carlson,Jon Copeland, J. W. Hastings, and the rest of the fireflycommunity for decades of sharing data, interpretationsand general insights. Helmut Hirsch has provided help,insight and critical evaluation of ideas. And finally, wethank a host of colleagues—John Buck, Thomas Eis-ner, Helmut Hirsch, J. W. Hastings, Su Tieman, andour anonymous ‘‘Reviewer No. 1’’—for their carefulreview of the manuscript and for their thoughtful andconstructive suggestions. Supported in part by NSFgrant PCM 78-22924 and SUNY Research Foundationgrant 320–7328A, both to H. Gh.

REFERENCES

Abele, D. 2002. The radical life giver. Nature 420:27.Aprille, J. R., C. J. Lagace, S. M. Lewis, T. Michel, J. S. Modica-

Napolitano, B. A. Trimmer, and R. M. Zayas. 2002. Mechanismof firefly flash control: Nitric oxide inhibition of oxygen con-sumption in lantern mitochondria is reversed by light. In P. E.Stanley and L. J. Kricka (eds.), Bioluminescence and chemilu-minescence: progress and current applications, pp. 25–28.World Scientific Publishing, Singapore.

Aprille, J. R., C. J. Lagace, J. Modica-Napolitano, and B. A. Trim-mer. 2004. Role of nitric oxide and mitochondria in control offirefly flash. Integr. Comp. Biol. 44:000–000.

Brau, F., J.-C. Bernengo, K.-L. Min, and J.-P. Steghens. 2000. Fireflyluciferase generates two low-molecular weight light-emittingspecies. Biochem. Biophys. Res. Commun. 270:247–253.

Brown, G. C. 1995. Reversible binding and inhibition of catalase bynitric oxide. Eur. J. Biochem. 232:188–191.

Buck, J. B. 1948. The anatomy and physiology of the light organin fireflies. Ann. N.Y. Acad. Sci. 49:397–482.

Carlson, A. D. 1965. Factors affecting firefly larval luminescence.Biol. Bull. 129:234–243.

Carlson, A. D. 2003. Photuris photocytes may differ among species.Symposium, Fireflies at Fifty, SICB Conference, Toronto, Can-ada, January 5. 2003.

Case, J. F. and L. G. Strause. 1978. Neurally controlled luminescentsystems. In P. J. Herring (ed.), Bioluminescence in action, pp.331–366. Academic Press, London.

Dahlgren, U. 1917. The production of light by animals. J. FranklinInst. 183:593–624.

Dow



arch 2022


de Duve, C. 1969. Evolution of the peroxisome. Ann. N.Y. Acad.Sci. 168:369–381.

DeLuca, M. and W. D. McElroy. 1974. Kinetics of the firefly lucif-erase catalyzed reactions. Biochemistry 13:921–925.

Ghiradella, H. 1977. Fine structure of the tracheoles of the lanternof a photurid firefly. J. Morphol. 153:187–204.

Ghiradella, H. 1998. The anatomy of light production: The finestructure of the firefly lantern. In M. Locke (ed.), F. M. Harrison(series ed.), Microscopic anatomy of invertebrates, 11A, Insecta,pp. 363–381. Wiley-Liss, New York.

Ghiradella, H. 2003. The mystery of flash control. Symposium, Fire-flies at Fifty, SICB Conference, Toronto, Canada, January 5.2003.

Ghiradella, H. 2004. Flash control in the firefly. In J. Capinera (ed.),Encyclopedia of entomology. Kluwer, Dordrecht, Netherlands.(In press)

Greenfield, M. 2001. Missing link in firefly bioluminescence re-vealed: NO regulation of photocyte respiration. BioEssays 23:992–995.

Hastings, J. W. and J. Buck. 1956. The firefly pseudoflash in relationto photogenic control. Biol. Bull. 111:101–113.

Hoffman, R. 2004. The story of O. Am. Sci. 92:23–26.Lane, N. 2002. Oxygen: The molecule that made the world. Oxford

University Press, Oxford.Lerner, R. A. and A. Eschenmoser. 2003. Ozone in biology. Proc.

Natl. Acad. Sci. U.S.A. 100:3013–3015.Maloeuf, N. S. R. 1938. The basis of the rhythmic flashing of the

firefly. Ann. Entomol. Soc. Amer. 31:374–380.McCapra, F. 1970. Chemiluminescence of organic compounds. Pure

Appl. Chem. 23:611–629.McElroy, W. D. and J. W. Hastings. 1956. Initiation and control of

firefly luminescence. In C. L. Prosser (ed.), Physiological trig-gers, pp. 80–84. Ronald Press, New York.

McElroy, W. D. and M. DeLuca. 1978. Chemistry of firefly lumi-nescence. In P. J. Herring (ed.), Bioluminescence in action, pp.109–127. Academic Press, New York.

Min, K.-L. and J.-P. Steghens. 1999. The emitting species dissociated

from the enzyme can emit the light in Photinus pyralis lucif-erase system. Biochem. Biophys. Res. Commun. 265:273–278.

Nathanson, J. A. 1979. Octopamine receptors, adenosine 39,59-monophosphate, and neural control of firefly flashing. Science203:65–68.

Nishiyama, M., A. Hoshino, L. Tsai, J. R. Henley, Y. Goshima, M.Tessier-Lavigne, M. M. Poo, and K. Hong. 2003. Cyclic AMP/GMP-dependent modulation of Ca21 channels sets the polarityof nerve growth-cone turning. Nature 423:990–995.

Oertel, D. and J. F. Case. 1976. Neural excitation of the larval fireflyphotocyte: Slow depolarization possibly mediated by a cyclicnucleotide. J. Exp. Biol. 65:213–227.

Oertel, D., K. A. Linberg, and J. F. Case. 1975. Ultrastructure of thelarval firefly light organ as related to control of light emission.Cell Tiss. Res. 164:27–44.

Smith, D. S. 1963. The organization and innervation of the lumi-nescent organ of the firefly, Photuris pennsylvanica (Coleop-tera). J. Cell Biol. 16:323–359.

Smith, D. S. 1968. Insect cells: Their structure and function. Oliverand Boyd, Edinburgh, UK.

Timmins, G. S., F. J. Robb, C. M. Wilmot, S. K. Jackson, and H.M. Swartz. 2001. Firefly flashing is controlled by gating oxygento light-emitting cells. J. Exp. Biol. 204:2795–2801.

Trimmer, B. A., J. R. Aprille, D. M. Dudzinski, C. J. Lagace, S. M.Lewis, T. Michel, S. Qazi, and R. M. Zayas. 2001. Nitric oxideand the control of firefly flashing. Science 292:2486–2488.

Wentworth, P., A. D. Wentworth, X. Zhu, I. A. Wilson, K. D. Janda,A. Eschenmoser, and R. A. Lerner. 2003. Evidence for the pro-duction of trioxygen species during antibody-catalyzed chemi-cal modification of antigens. Proc. Natl. Acad. Sci. U.S.A. 100:1490–1493.

Wigglesworth, V. B. 1983. The physiology of insect tracheoles. InM. J. Berridge, J. E. Treherne, and V. B. Wigglesworth (eds.),Advances in insect physiology, Vol. 17, pp. 85–148. AcademicPress, New York.

Wilson, T. and J. W. Hastings. 1998. Bioluminescence. Annu. Rev.Cell Dev. Biol. 14:197–230.

Zeng, H., B. G. Loughton, and K. R. Jennings. 1996. Tissue specifictransduction systems for octopamine in the locust (Locusta mig-ratoria). J. Insect Physiol. 42:765–769.

Dow



arch 2022

Fireflies at One Hundred Plus: A New Look at Flash Control1

Documents

Transcript of Fireflies at One Hundred Plus: A New Look at Flash Control1