Distinct evolutionary trajectories of two integration centres, the central complex and mushroom bodies, across Heliconiini butterflies

Reviewer #1 (Public review):

The authors previously reported that Heliconius, one genus of the Heliconiini butterflies, evolved to be efficient foragers to feed pollen of specific plants and have massively expanded mushroom bodies. Using the same image dataset, the authors segmented the central complex and associated brain regions and found that the volume of the central complex relative to the rest of the brain is largely conserved across the Heliconiini butterflies. By performing immunostaining to label a specific subset of neurons, the authors found several potential sites of evolutionary divergence in the central complex neural circuits, including the number of GABAergic ellipsoid body ring neurons and the innervation patterns of Allatostatin A expressing neurons in the noduli. These neuroanatomical data will be helpful to guide future studies to understand the evolution of the neural circuits for vector-based navigation.

Strengths:

The authors used a sufficiently large scale of dataset from 307 individuals of 41 species of Heliconiini butterflies to solidify the quantitative conclusions and present new microscopy data for fine neuroanatomical comparison of the central complex.

Weaknesses:

(1) Although the figures display a concise summary of anatomical findings, it would be difficult for non-experts to learn from this manuscript to identify the same neuronal processes in the raw confocal stacks. It would be helpful to have instructive movies to show a step-by-step guide for identification of neurons of interest, segmentations, and 3D visualizations (rotation) for several examples, including ER neurons (to supplement texts in line 347-353) and Allatostatin A neurons.

(2) Related to (1), it was difficult for me to assess if the data in Figure 7 support the author's conclusions that ER neuron number increased in Heliconius Melpomene. By my understanding, the resolution of this dataset isn't high enough to trace individual axons and therefore authors do not rule out that the portion of "ER ring neurons" in Heliconius may not innervate the ER, as stated in Line 635 "Importantly, we also found that some ER neurons bypass the ellipsoid body and give rise to dense branches within distinct layers in the fan-shaped body (ER-FB)". If they don't innervate the ellipsoid body, why are they named as "ER neurons"?

(3) Discussions around the lines 577-584 require the assumption that each ellipsoid body (EB) ring neuron typically arborises in a single microglomerulus to form a largely one-to-one connection with TuBu neurons within the bulb (BU), and therefore, the number of BU microglomeruli should provide an estimation of the number of ER neurons. Explain this key assumption or provide an alternative explanation.

(4) The details of antibody information are missing in the Key resource table. Instead of citing papers, list the catalogue numbers and identifier for commercially available antibodies, and describe the antigen, and whether they are monoclonal or polyclonal. Are antigens conserved across species?

(5) I did not understand why authors assume that foraging to feed on pollens is a more difficult cognitive task than foraging to feed on nectar. Would it be possible that they are equally demanding tasks, but pollen feeding allows Heliconius to pass more proteins and nucleic acids to their offspring and therefore they can develop larger mushroom bodies?

Reviewer #2 (Public review):

Summary:

In this study, Farnsworth et al. ask whether the previously established expansion of mushroom bodies in the pollen foraging Heliconius genus of Heliconiini butterflies co-evolved with adaptations in the central complex. Heliconius trap line foraging strategies to acquire pollen as a novel resource require advanced spatial memory mediated by larger mushroom bodies, but the authors show that related navigation circuits in the central complex are highly conserved across the Heliconiini tribe, with a few interesting exceptions. Using general immunohistochemical stains and 3D reconstruction, the authors compared volumes of central complex regions, and unlike the mushroom bodies, there was no evidence of expansion associated with pollen feeding. However, a second dataset of neuromodulator and neuropeptide antibody labeling reveals more subtle differences between pollen and non-pollen foragers and highlights sub-circuits that may mediate species-specific differences in behavior. Specifically, the authors found an expansion of GABAergic ER neurons projecting to the fan-shaped body in Heliconius, which may enhance their ability to path-integrate. They also found differences in Allatostatin A immunoreactivity, particularly increased expression in the noduli associated with pollen feeding. These differences warrant closer examination in future studies to determine their functional implication on navigation and foraging behaviors.

Strengths:

The authors leveraged a large morphological data set from the Heliconiini to achieve excellent phylogenetic coverage across the tribe with 41 species represented. Their high-quality histology resolves anatomical details to the level of specific, identifiable tracts and cell body clusters. They revealed differences at a circuit level, which would not be obvious from a volumetric comparison. The discussion of these adaptations in the context of central complex models is useful for generating new hypotheses for future studies on the function of ER-FB neurons and the role of Allatostatin A modulation in navigation.

The conclusions drawn in this paper are measured and supported by rigorous statistics and evidence from micrographs.

Weaknesses:

The majority of results in this study do not reveal adaptations in the central complex associated with pollen foraging. However, reporting conserved traits is useful and illustrates where developmental or functional constraints may be acting. The implied hypothesis in the introduction is that expansion of mushroom bodies in Heliconius co-evolved with central complex adaptations, so it may be helpful to set up the alternate hypotheses in the beginning.

In the main text, the authors describe differences in GABAergic neurons "across several species" but only one Heliconius and one outgroup species seem to be represented in the figures. ER numbers in Figure 7H are only compared for these two species. If this data is available for other species, it would strengthen the paper to add them to the analysis, since this was one of the most intriguing findings in the study. I would want to know if the increased ER number is a trend in Heliconius or specific to H. melpomene.

Author response:

We thank the two reviewers for their constructive criticisms which we will address in the coming weeks, and we are confident doing so will benefit the manuscript.

We will aim to address all comments, but there are two main areas in particular that we highlight here:

(1) Both reviewers make important suggestions to improve the readers’ understanding of the anatomical complexities and raw files we provide. We will generate annotated confocal stacks and simplify the nomenclature to better guide the reader through the more complex details of the anatomy of the central complex, and the neuron types we characterized more closely.

(2) Both reviewers also pointed to several parts of our interpretations and discussion that should be clarified. We will do so by improving the language we use at certain sections to offer more precision, and by offering alternative explanations where possible.