Module 6: Monarch Compass & Biochemistry

Steven Reppert’s group at the University of Massachusetts Medical School established the monarch as the premier insect navigation model over a fifteen-year programme culminating in the 2009 demonstration that the circadian pacemaker driving the sun compass is located in the antennae, not in the brain (Merlin, Gegear & Reppert, Science, 2009). This was genuinely surprising: in Drosophila and most other insects the master clock resides in a small set of lateral neurons of the brain. In migrating monarchs the antennal clock over-rides the brain clock for the purpose of azimuthal time-compensation.

The experimental design is disarmingly simple. Merlin painted the antennae of test butterflies with black enamel to block all photic input, while leaving the compound eyes and brain-DN clock fully functional. In a flight simulator the black-antenna monarchs failed to maintain the south-south-west migratory heading—they wandered. The control groups (clear enamel, or painted thorax) navigated normally. When Merlin subjected animals to a 6-hour phase-shifted light-dark cycle for six days (‘jet-lagging’ the antennal clock), the heading rotated precisely by the expected 90° (6 h × 15 °/h). Blocking photic input only to the eyes, with antennae exposed, did not produce this effect.

Earlier work leading to the result

Before the antennal discovery, Sauman et al. (Neuron, 2005) had identified the period (per), timeless (tim) and cryptochrome-1 (cry1) transcripts in the monarch brain, and Heinze & Reppert (Neuron, 2011) had mapped the central-complex integrator neurons that combine skylight polarisation, chromatic, and intensity cues with azimuth. The antennal-clock result of 2009 shifted the paradigm: the brain clock still sets activity rhythms, but the phase reference for time-compensation of the sun-compass is in the antennae.

Gustav Kramer’s 1950 starling experiments introduced the concept of the time-compensated sun compass: to hold a fixed geographical bearing a diurnal migrant must rotate its reference azimuth at 15 °/h to track the sun’s apparent motion. The internal clock supplies the ‘hour hand’. Formally, if \(A_\odot(t)\) is the solar azimuth at local time t and \(\phi(t_\mathrm{internal})\) is the expected azimuth for south based on the internal clock reading, the inferred compass heading is

When the internal clock is phase-shifted by \(\Delta t\) relative to solar time, the resulting heading error is \(\Delta \hat\theta \approx \omega_\oplus\, \Delta t\)where \(\omega_\oplus = 15\,\)°/h is the Earth’s rotation rate. A clock shift of +6 h produces a heading drift of +90°, which is precisely what Merlin measured in black-enamel vs antenna-jet-lagged monarchs. The quantitative match is a hallmark of a true time-compensation mechanism.

Solar ephemeris

At latitude \(\varphi\) and declination \(\delta\), the solar azimuth is obtained from the spherical-astronomy identity

The core molecular oscillator of the monarch antennal clock is broadly conserved with Drosophila. Two transcription factors, CLOCK (CLK) and CYCLE (CYC), heterodimerise and bind E-box promoter elements to drive transcription of the repressor genes period (per) and timeless(tim). PER and TIM proteins accumulate in the cytoplasm, gradually migrate to the nucleus, and inhibit CLK:CYC—closing a negative feedback loop with roughly 24 h period.

Dual cryptochrome system

Unlike the Drosophila model, the monarch carries both insect-like (type 1) and vertebrate-like (type 2) cryptochromes. CRY1 is light-responsive, absorbing UV-blue light (\(\lambda \approx 350{-}450\,\)nm) and participating in TIM degradation as in fly. CRY2, by contrast, is light-independent and acts as a transcriptional repressor, more akin to the mammalian cryptochromes. Zhu et al. (PLoS Biol., 2008) cloned both and showed that their expression patterns in the monarch antenna and brain are distinct but overlapping.

Why the antenna?

The antenna is a peripheral sensory organ directly exposed to environmental light, temperature, and humidity. Locating the time-compensation clock there may permit rapid entrainment to local dawn/dusk photoperiod cues, which would be dampened by filtering through the head capsule. In the super-generation monarch the antennal clock is also where diapause-hormone signalling interacts with the photoperiodic history, integrating seasonal and daily time information.

Stanley Heinze and Steven Reppert traced the sun-compass signal from retina to the central complex, the insect analogue of vertebrate basal ganglia plus cerebellum. Polarisation-sensitive photoreceptors in the dorsal rim area of the compound eye project to the medulla, then through the anterior optic tubercle (AOTu) to the bulb and the protocerebral bridge. Within the central body the signal is mapped onto a ring of compass neurons whose activity encodes the current heading. The same neurons receive solar azimuth and polarisation-pattern input, and fire in a manner directly analogous to the head-direction cells of the mammalian limbic system.

Cloudy days and dense forest canopies deny the butterfly its sun compass for hours at a time. Guerra, Gegear & Reppert (Nature Communications, 2014) showed that monarchs also carry a light-dependent magnetic inclination compass. In a circular flight-simulator arena with no visible celestial cues, monarchs held a consistent southward orientation when the simulated geomagnetic inclination pointed equatorward, and reversed orientation when the inclination was flipped by Helmholtz coils. Blocking UV-A light (\(\lambda < 420\,\)nm) abolished the response, implicating a light-dependent radical-pair photoreceptor (see Module 2 for the general radical-pair theory).

Mechanism debate

The molecular identity of the monarch magnetoreceptor is still debated. Both CRY1 and CRY2 are candidates on the basis of UV-blue sensitivity of CRY1 and radical-pair chemistry of the conserved FAD-binding pocket. Neither has been proven to give rise to the behavioural response, however. Bazalova et al. (PNAS, 2016) argued from CRY1-knockout experiments in related insects that CRY1 may not be essential, suggesting CRY2 or an unidentified receptor. Others have noted that biogenic-magnetite chains found in some monarch tissues might support a complementary force-based mechanism. The question remains open.

The monarch caterpillar feeds almost exclusively from milkweed (Asclepias) and from the related Cynanchum, Matelea, and dogbane genera in the Apocynaceae. Milkweed latex is rich in cardenolides (cardiac glycosides), a family of steroidal natural products that specifically inhibit the Na⁺/K⁺-ATPase of animal cells. The caterpillar consumes, does not metabolise, and sequesters these toxins; the adult retains 30–80% of the larval cardenolide load, primarily in wing scales and cuticle.

Cardenolide diversity

Over 200 distinct cardenolides have been identified in Asclepias, including aspecioside, calotropin, calactin, uscharin, and syrioside. Species vary enormously in total concentration: A. syriaca (common milkweed, the dominant midwestern host) bears roughly 400 μg/g leaf dry mass; A. humistrata 2 500 μg/g; and the invasive tropical A. curassavica 3 000–5 000 μg/g. Monarchs reared on higher-cardenolide species accumulate more toxin but suffer more larval mortality (Malcolm 1994). The relationship between ingested and sequestered cardenolide is approximately linear at low concentrations and saturates at the highest doses as sequestration machinery is overloaded.

Cardenolides bind in a cleft formed between the M1 and M2 transmembrane helices of the Na⁺/K⁺-ATPase alpha-subunit, with a mammalian K_i in the range of 10⁻⁸ M. The binding pocket is lined by residues whose identities determine ouabain sensitivity. Dobler, Dalla, Wagschal & Agrawal (PNAS, 2012) sequenced the ATP1A1 alpha-subunit across 21 arthropod species adapted to cardenolide host plants and identified three recurrent amino-acid substitutions in the H1-H2 extracellular loop: Q111V, A119S, and N122H. All three are present in monarchs.

Karageorgi 2019: CRISPR reconstruction

Marianthi Karageorgi, Noah Whiteman and colleagues (Nature, 2019) used CRISPR-Cas9 to install each of the three monarch substitutions, alone and in combination, into the Drosophila melanogaster ATP1A1 orthologue. The single substitution N122H conferred ~15-fold resistance; the double N122H + A119S ~100-fold; and the full triple Q111V + A119S + N122H ~150-fold resistance—closely matching native monarch enzyme kinetics. Strikingly, the intermediate genotypes showed epistasis: the A119S substitution alone produced pleiotropic motor deficits in fly larvae, which were rescued by the addition of N122H. This maps the evolutionary sequence: N122H appears first, rescuing A119S pleiotropy, with Q111V fine-tuning resistance last. The work reconstructs an adaptive landscape in a non-model insect.

Lincoln Brower’s 1968 Scientific American paper, ‘Ecological Chemistry’, documented for the first time that monarch cardenolide load translates directly into vertebrate predator rejection. Brower, Van Brower & Corvino fed captive blue jays (Cyanocitta cristata) monarchs reared on A. curassavica (high-cardenolide) or on non-toxic cabbage plants (low-cardenolide control). The cardenolide-reared butterflies induced vomiting within 12–15 minutes in all trial birds. After a single emetic experience, the jays refused all subsequent orange-and-black butterflies, monarch or cabbage-reared, for up to four weeks.

Müllerian mimicry with the viceroy

The viceroy butterfly (Limenitis archippus) was long thought to be a Batesian (undefended) mimic of the monarch. Ritland & Brower (Nature, 1991) revised this century-old claim, demonstrating that the viceroy itself is unpalatable—it sequesters salicylic-acid derivatives from willow—and that it mimics monarchs in the Müllerian sense (both models share the cost of educating predators). The viceroy and monarch therefore occupy a shared aposematic signal.

Cardenolide-tolerant predators

Some species tolerate moderate cardenolide doses and prey on the overwintering Mexican monarch clusters: the black-headed grosbeak (Pheucticus melanocephalus) and the black-backed oriole (Icterus abeillei) can each consume 100–200 monarchs per day by excising the cardenolide-rich abdomen and wing muscles. Arellano-Guillermo et al. (1993) estimated that these two species collectively remove up to 10% of the overwintering cohort in some years. Black-eared mice (Peromyscus melanotis) at the overwintering sites also scavenge fallen monarchs at night.

Where did any given Mexican-overwintering monarch originate? Keith Hobson, Leonard Wassenaar & Orley Taylor (Oecologia, 1999) developed a stable-isotope-based provenance assay using the hydrogen and oxygen isotope composition of wing chitin, \(\delta^2H_\mathrm{wing}\) and \(\delta^{18}O_\mathrm{wing}\). Continental-scale hydrogen isoscape maps show a steep latitudinal gradient in precipitation \(\delta^2H\) (lighter values at higher latitudes), and this signal is transmitted through milkweed tissue into the caterpillar and fixed in the wing chitin laid down at pupation. Sampling overwintering monarchs allows the natal latitude to be reconstructed to within a few hundred kilometres.

Wassenaar & Hobson (PNAS, 1998) analysed 597 overwintering monarchs from seven Mexican colonies and showed that the midwestern US corn belt contributes the majority (50–80%) of overwintering individuals, with smaller contributions from the northeastern US and Canadian Prairies. Flockhart et al. (J. Animal Ecology, 2015) extended the analysis and showed that the midwestern fraction has remained roughly constant even as total numbers declined, confirming that midwestern milkweed loss is the dominant driver of the population collapse.

Monarch orange-and-black wing colouration is among the most-studied cases of aposematism (warning colouration) in insects. The orange pigment is derived from papiliochrome-type ommochromes and 3-hydroxykynurenine; the black is melanised scales. The high-contrast pattern is visible to avian predators at distances of tens of metres and is memorised after a single unpalatable experience (Brower 1968; Ritland & Brower 1991). The ommochromes absorb strongly in the UV as well as the visible, potentially contributing to UV-wavelength recognition by conspecifics.

Even when the sun itself is hidden by clouds, the sky retains a characteristic e-vector polarisation pattern determined by single Rayleigh scattering of unpolarised sunlight in the atmosphere. The degree of polarisation peaks at 90° from the sun, tracing a great circle perpendicular to the solar vector, and the electric-field vector at any point in the sky is perpendicular to the plane defined by that point, the observer, and the sun. Karl von Frisch demonstrated in the 1940s that honeybees use this polarisation pattern as a sun compass back-up, and subsequent work in Drosophila, desert ants, locusts, and crickets established it as a general invertebrate navigation cue.

In monarchs the polarisation-sensitive photoreceptors occupy the dorsal rim area (DRA) of the compound eye. DRA ommatidia have rhabdomeres with aligned microvilli, making them intrinsically dichroic. The dichroic absorption tunes the receptor to respond maximally to linearly polarised light with a specific e-vector direction. Population-level downstream processing combines the signals from across the DRA to extract a single solar-azimuth estimate, even when the sun itself is invisible. Reppert et al. (PLoS ONE, 2004) and Heinze & Reppert (2011) showed that the monarch central complex receives and integrates both DRA polarisation input and direct solar-azimuth input via the chromatic channel.

The monarch thus carries not a single compass but a layered redundant system: sun azimuth (primary on clear days), skylight polarisation (secondary when sun is obscured), and the magnetic inclination compass (tertiary, active when both visual cues are unavailable). Each compass has limited precision—a few degrees for the sun, about 10° for polarisation, and 20–30° for the magnetic system. Combined, the three channels can produce a composite heading estimate with accuracy well below 5° across a wide range of conditions.

The combining algorithm is a Bayesian weighted average: when sensor variance is low the corresponding channel dominates; when variance is high (clouds, magnetic storms) the channel is downweighted automatically. This kind of multi-sensor fusion is mathematically equivalent to a Kalman filter over heading. The central complex implements it approximately through gain modulation at each input stage, and the algorithm is robust to failure of any single channel—a property that becomes essential on the 60-day migration through variable weather.

The cardenolide-resistance substitutions in ATP1A1 are found not only in monarchs but across dozens of lineages of milkweed-feeding insects: large milkweed bugs (Oncopeltus fasciatus), dogbane beetles (Chrysochus auratus), African lubber grasshoppers, queen butterflies (Danaus gilippus), and several genera of danaine butterflies. Dobler et al. (2012) counted at least 21 independent origins of the N122H substitution across arthropods, a striking case of parallel evolution at a single amino-acid site. The underlying reason is the structural constraint of the Na⁺/K⁺-ATPase binding pocket: only a very small set of substitutions can both block cardenolide binding and preserve the ion-pumping function.

Petschenka et al. (Mol. Biol. Evol., 2017) added further depth by showing that the triple substitution in monarchs is present in multiple paralogous alpha-subunit genes (ATP1A1, ATP1A2, ATP1A3) arranged along the genome as tandem duplicates. The duplicates are differentially expressed in different tissues: gut ATPase has the full triple resistance, whereas nerve-tissue ATPase has only N122H. This tissue-specific expression pattern reflects a resource-allocation trade-off: full resistance is metabolically costly and is deployed only where cardenolide exposure is greatest.

A side consequence of home-garden planting of exotic tropical milkweed (Asclepias curassavica) in the southern US is the emergence of year-round monarch breeding populations that no longer migrate. Satterfield et al. (Proc. R. Soc. B, 2015) showed that non-migratory monarchs on A. curassavica have 5–15× higher OE parasite loads, reduced reproductive output, and impaired flight ability. The tropical milkweed also retains its leaves year-round in the US South, providing continuous larval habitat and breaking the normal migratory diapause cue. The conservation implication is direct: planting exotic tropical milkweed in the southern US may do more harm than good, and native milkweed species that senesce with natural seasonality are a better choice.

A super-generation monarch in late September embarks on a 4 000 km migration using an antennal clock to time-compensate a sun compass, a UV-sensitive cryptochrome-mediated magnetic back-up on overcast days, and a central-complex ring network to integrate the compass cues into a steering command. Throughout the journey its body is armoured chemically with cardenolides sequestered from the milkweed of its larval host, rendering it unpalatable to all but the most cardenolide-tolerant avian predators. Its own Na⁺/K⁺-ATPase tolerates the sequestered toxin thanks to three amino-acid substitutions produced by tens of millions of years of coevolution with its host. The integrated picture is one of the most tightly coupled sensor-chemistry-ecology systems in all of animal biology.

Module 7 next addresses the energetic side of the picture: how the super-generation accumulates and burns fuel to make the long-haul flight possible, and how its body plan is remodelled during hyperphagia.