Circuit convergence, divergence, and re-convergence can be found throughout the nervous systems of both invertebrates and vertebrates and play a pivotal role in providing behavioral flexibility (Eschbach et al., 2020; Jeanne and Wilson, 2015; Man et al., 2013; Miroschnikow et al., 2018; Mišić et al., 2014; Ohyama et al., 2015). Given the importance of the MBONs in driving behavioral choice, we first sought to reveal patterns of divergence and convergence by identifying the postsynaptic connections of the MBONs innervating each of the 15 MB compartments using trans-Tango (Talay et al., 2017). Since trans-Tango signal depends on the strength and specificity of the GAL4 driver being used, we selected 28 previously published MBON split-GAL4 lines specific to individual MBONs, or sparse but overlapping subsets of MBONs (Aso et al., 2014a). We combined trans-Tango with chemogenetic active zone marker using the brp-SNAP knock-in to increase uniformity of neuropil labeling (Kohl et al., 2014).

We successfully identified the postsynaptic connections of 25 split-GAL4 lines (Figure 1, Figure 1—figure supplements 1–23, open access raw data video files are available at https://doi.org/10.26300/mttr-r782). trans-Tango signals from MB112C (MBON γ1pedc>α/β) and G0239 (MBON α3) were too weak and were excluded from further analysis. In contrast, signals from MB242A (MBON calyx) proved to be too noisy to confidently identify postsynaptic connections. We also employed three new split-GAL4 lines that had more specific expression for γ5β′2a, β′2mp, and α2sc MBONs. Postsynaptic connections of glutamatergic, GABAergic, and cholinergic MBONs vary with regard to the divergence and breadth of their postsynaptic connections (Figure 1, Figure 1—figure supplements 1–23, external open access raw data video files are available at https://doi.org/10.26300/mttr-r782). For instance, MB011B, which includes glutamatergic MBONs γ5β′2a, β′2mp, and β′2mp-bilateral has extensive connections across the superior protocerebrum (Figure 1A), whereas MB542B, which includes cholinergic MBONs α′1, α2p3p, α′3 m has limited connections within the LH (Figure 1N). The innervation patterns did not seem to correlate with neurotransmitter type or number of MBONs expressing each split-GAL4.

However, it was clear that some of the data were confounded by split-GAL4 lines that had off-target expression. We excluded extraneous signals by segmenting trans-Tango signals that were continuous with MBON terminals (Figure 2A–B) and then quantified the distribution of postsynaptic signals across brain regions in the standard brain (Ito et al., 2014). Nearly all MBONs have divergent connections across the dorsal brain regions, CRE, SMP, SIP, SLP, LH, as well as FSB, and LAL (Figure 2C–E, Figure 2—figure supplement 1).

Of the DANs innervating the MB, 90% have dendritic arborizations that are localized to four of the five proposed MBON convergent regions, including CRE, SMP, SIP, and SLP (Aso et al., 2014a). Subsets of MBON axons overlapping with DAN dendritic arborizations provide feedback opportunities for MBONs to modulate DAN input thereby indirectly modulating MB circuits. Thus, we selected a subset of MBONs that were reported to co-localize with protocerebral anterior medial (PAM) DANs and co-stained with antibodies against tyrosine hydroxylase (TH) to identify overlap with trans-Tango signal (Aso et al., 2014a). As expected, some of the neurons postsynaptic to MBONs were TH positive; however, due to the complexity of trans-Tango-labeled neurons, we were unable to identify the DANs postsynaptic to a particular MBON unequivocally. Most overlap between TH and trans-Tango signals was observed with γ5β′2a (MB011B, 25 ± 0.7; n = 4; Figure 3A) and β′2mp (MB002B, 10.25 ± 1.3 n = 4 and MB074C, 4.75 ± 1.1, n = 4; Figure 3B and C) MBONs. These MBONs were predicted to co-localize with PAM DANs β′2p, β′2m and PAM DANs γ5 and β′2a, respectively (Aso et al., 2014a). Similarly, the γ3, γ3β′1 MBON was predicted to overlap with PAM γ3 and β′1m, and MB083C had an average of nine cells (9 ± 2.0, n = 10) with co-expression of TH and trans-Tango signals (Figure 3D). Likewise, the cholinergic γ2α′1 MBON (MB077C) was predicted to overlap with PAM γ4>γ1γ2, and indeed, MB077C brains averaged five cells (5 ± 1.5, n = 8) with co-expression of TH and trans-Tango signals per hemibrain (Figure 3E). There were a number of MBONs that had very few or no TH-positive postsynaptic neurons (Figure 3—figure supplement 1). The majority of these MBONs innervate the vertical lobe, including MBON α1 (MB310C; Figure 3—figure supplement 1A), MBON α′3ap, α′3 m (MB027B; Figure 3—figure supplement 1E), MBON α2sc (MB080C; Figure 3—figure supplement 1F) and MBON α′1, α2p3p, α′3m (MB542B; Figure 3—figure supplement 1G). MBON β′1 also had limited TH-positive postsynaptic neurons (MB057B; Figure 3—figure supplement 1D). Despite predictions that γ4>γ1γ2 MBON (MB298B) would co-localize with PAM γ4>γ1γ2, we found minimal co-expression of TH and trans-Tango signals (Figure 3—figure supplement 1C). This is likely a false negative due to the strength of the driver as annotations of the EM data has revealed postsynaptic connections with PAM γ4>γ1γ2 (Clements et al., 2020; Li et al., 2020). It is possible that the number of co-localized TH+ cells in our analysis here is an underestimation since some of the brains had fewer than expected TH+ neurons (Figure 3—figure supplement 2).

Whole brain overlap analysis identified the MB itself as a site of rich convergence for most MBON lines (Figure 2C). MBON postsynaptic signals in MB were not surprising given that many MBONs provide feedforward connections between MB compartments (Aso et al., 2014a). For instance, MBON γ4>γ1γ2 has dendritic arbors in γ4 and axonal projections in γ1γ2, MBON γ1pedc>α/β have dendritic arbors in γ1 and axonal projections in α/β lobes, and MBON β1>α has dendritic arbors in β1 and axon projections to the entire alpha lobe. However, further analysis revealed that in addition to providing connections between MB compartments, MBONs converge directly on other MBONs presumably through axo-axonal connections. Two different MBONs are frequently targeted: MBON β′2mp (Figure 4A) and MBON γ3β′1 (Figure 4B). Interestingly, MBON β′2mp receives convergent glutamatergic, GABAergic, and cholinergic input from MBON γ5β′2a (MB011B and MB210B), MBON γ3β′1 (MB110C and MB83C), MBON α′2 (MB018B and MB082C), and MBON γ2α′1 (MB077B and MB051C) (Figure 4A, Figure 4—figure supplement 1). MBON γ3β′1 receives convergent input from glutamatergic MBON β′2mp as revealed with split-GAL4 lines MB002B (Figure 4B) and MB074C (Figure 4—figure supplement 1) and glutamatergic MBON γ4>γ1γ2 (MB298B, Figure 4B). We hypothesize that similar to MBONs that project to other regions of the MB, MBON γ3β′1, and MBON β′2mp create opportunities for multilevel feedforward networks to update information to drive behavioral response (Figure 4C).

Another site of convergence of the MBON network was the FSB (Figure 5). MBON postsynaptic connections display a laminar organization primarily across the dorsal region of the FSB. Nearly all the glutamatergic and GABAergic MBONs converge onto FSB layers 4 and 5, and to a lesser extent, layer 6 (Figure 5A–B). MBON α1 is the only type of MBON that had broad trans-Tango signals in the FSB (Figure 5A). To rule out sexual dimorphism in the postsynaptic connections of MBON α1, we compared trans-Tango signal in the FSB in male and female brains and found similar innervation patterns (Figure 5—figure supplement 1). Cholinergic MBONs also had trans-Tango signals in the dorsal FSB but with more variability across MBON lines and within each line (Figure 5C). For instance, trans-Tango with MBON γ2α′1 consistently visualized projections to FSB layers 4 and 5 in all the brains analyzed, whereas more variability was observed in FSB innervation pattern across MBON α′2 brains (Figure 5—figure supplement 2). MBON α′1 and α2sc both project exclusively to FSB layer 6 (Figure 5C). Together, FSB layers 4 and 5 receive convergent input from combinations of glutamatergic, GABAergic and cholinergic MBONs (Figure 5D; Figure 5E).

Both visual and computational analyses confirmed the CRE, SMP, SIP, and SLP, as well as the MB and FSB as obvious postsynaptic targets of the MBON network. Visual inspection also confirmed the LAL as postsynaptic to multiple MBON lines. Its identification was less obvious in computational analysis largely because the neurites innervating the LAL were not as extensive as the LAL itself and were often difficult to segment. Although not extensive, LAL innervation was consistent across glutamatergic, GABAergic, and cholinergic MBONs (Figure 6). Specifically, glutamatergic γ5β′2a, β′2mp, and β′2mp_bilateral had postsynaptic neurites within the LAL in all of the brains analyzed (Figure 6A). Similarly, GABAergic MBON γ3, γ3β′1, and β′1 (Figure 6B) and cholinergic MBON γ2α′1 (Figure 6C) consistently had postsynaptic neurites within the LAL. Thus, like the FSB, neurons innervating the LAL receives convergent input from combinations of glutamatergic, GABAergic and cholinergic MBONs (Figure 6D; Figure 6E).

Thus far, we have confirmed two postsynaptic targets of the MBON network that reside outside of the MB: the FSB and LAL. However, the identities of the postsynaptic neurons within FSB and LAL as well as their functions remain unknown. Our strategy for identifying FSB and LAL neurons and interrogating their functional connectivity with MBONs was to selectively label neurons in FSB and LAL using specific drivers and to examine whether they are co-localized with postsynaptic signal when we initiate trans-Tango from MBONs. To achieve this, we identified candidate FSB and LAL LexA lines by performing a mask search of the LexA lines that have overlapping expression within the convergent region and brought them together with MBON lines: MB051C and MB077C were used to target MBON γ2α′1, MB083C and MB110C were used to target γ3β′1, and MB074C was used to target MBON β′2mp. We identified three candidate LexA lines: one to target FSB layer four neurons - R47H09 (Jenett et al., 2012; Pfeiffer et al., 2013; Pfeiffer et al., 2010), and two to target LAL neurons - VT055139 and VT018476 (Tirian and Dickson, 2017). Finally, we generated trans-Tango reporter flies where the UAS-myrGFP was replaced with UAS-CD2, and LexAOp-mCD8::GFP was included in order to visualize the starter MBONs, the postsynaptic trans-Tango signal, and the LexA lines simultaneously.

We successfully combined the majority of the targeted MBON split-Gal4 lines with FSB and LAL LexA lines (we were unable to combine MB074C with LexA line 47H09). Interestingly, for the cholinergic MBON γ2α′1 (MB077C), we identified at least two postsynaptic FSB neurons (labeled in the 47H09 LexA line; Figure 7A) and at least five postsynaptic LAL neurons (labeled in the VT055139 LexA line; Figure 7B). We next sought to interrogate functional connectivity between MBON γ2α′1 and 47H09 FSB neurons and VT055139 LAL neurons by combining optogenetic stimulation of MBON γ2α′1 using UAS-Chrimson and functional calcium imaging in FSB and LAL using LexAop-GCaMP6s. Stimulation of cholinergic MB077C with 400–500 ms of red light (627 nm) resulted in an increase in calcium signal in the FSB and LAL (Figure 7C). Similar activation of other cholinergic MBONs (MB080C), which do not innervate the LAL or layer 4 of the FSB, did not result in signal (Figure 7—figure supplement 1), supporting the specificity of this interaction and suggesting that the MBON γ2α′1 is both anatomically and functionally connected to the FSB and LAL. Strikingly, GABAergic MBON γ3β′1 (MB083C) also had at least one identified postsynaptic FSB neuron that was included in the expression of FSB 47H09 LexA line (Figure 7D) and at least two identified postsynaptic LAL neurons that were included in the expression of LAL VT055139 LexA line (Figure 7E). Thus, the genetically identified subsets of LAL and FSB neurons receive convergent input from GABAergic and cholinergic MBONs (Figure 7F). We hypothesize that the convergence of excitatory and inhibitory input onto both the LAL and FSB is critical for guiding behavior.

Finally, to determine the role of LAL neurons in the context of guiding behavior of flies in groups, we performed analyses of group activity using thermogenetic inactivation of identified split-GAL4 LAL neurons (Scaplen et al., 2019). Individual flies were tracked offline using Flytracker to obtain activity-based features (Eyjolfsdottir et al., 2014). Inactivation of SS32219-GAL4-positive LAL neurons (Figure 7G) resulted in significant increases in group activity (Figure 7H, (F(2,21)=39.28 p<0.0001), pathlength (F(2,29)=33.39, p<0.0001), angular velocity (F(2,29)=51.87, p<0.0001) and velocity (F(2,29)=30.97, p<0.0001) of individual flies (Figure 7I)). Behavioral results were replicated with a separate LAL split-GAL4 line (SS32230-GAL4, Figure 7—figure supplement 2), suggesting that LAL neurons downstream of MBONs modulate locomotor activity of flies in a group. Group activity at permissive temperatures was not different from controls (Figure 7—figure supplement 3).

Successive levels of convergence and divergence across the brain permit functional flexibility (Jeanne and Wilson, 2015; Man et al., 2013; Tye, 2018). Like the mushroom body, cerebellar circuits in mammals exhibit large divergence in connectivity, and this can support diverse types of synaptic plasticity (Litwin-Kumar et al., 2017). Previous neuroanatomical work in insects described divergent afferent and efferent MB neurons, although the extent of this divergence was unknown (Ito et al., 1998; Li and Strausfeld, 1997; Li and Strausfeld, 1999; Mao and Davis, 2009; Nässel and Elekes, 1992; Tanaka et al., 2008; Waddell, 2013). Our data revealed varying levels of divergence of postsynaptic connections of MBONs across the brain. Every one of the analyzed MBONs had postsynaptic partners projecting to multiple brain regions (Figure 2C, Figure 8A). Further, nearly the entire superior protocerebrum as well as portions of the inferior protocerebrum received input from at least one MBON, providing opportunities for comprehensive integration of signals from the MBON network.

Multiple feedforward and feedback circuits exist within the MB (Aso et al., 2014a; Eichler et al., 2017; Eschbach et al., 2020; Otto et al., 2020; Takemura et al., 2017; Zheng et al., 2018). Our data revealed at least two MBONs that receive convergent input from multiple MBONs and are also reciprocally connected (Figure 3, Figure 8B). The convergent MBON input to β′2mp is especially interesting as cholinergic (MBON γ2α′1), GABAergic (MBON γ3β′1), and glutamatergic (MBON γ5β′2a) MBONs drive opposing behaviors (Aso et al., 2014b). For instance, activation of the cholinergic or GABAergic MBON results in naive odor preference, whereas activation of the glutamatergic MBON results in robust naive avoidance (Aso et al., 2014b; Lewis et al., 2015). Similarly, the cholinergic MBON activity mediates aversive associations (Berry et al., 2018; Yamazaki et al., 2018), whereas glutamatergic MBON activity mediates appetitive associations and extinction of aversive memories (Felsenberg et al., 2018; Owald et al., 2015; Yamazaki et al., 2018).

Considering that MBON β′2mp receives convergent input from these parallel and opposing pathways (Figure 4C), it likely serves as a decision hub by integrating activity to modulate cue-induced approach and avoidance behavior. How MBON β′2mp integrates information across MBONs and drives behavioral responses remains to be determined. Naïve activation of MBON β′2mp does not appear to influence behavioral choice, it instead acts as a sleep suppressor (Aso et al., 2014b). Inhibition of MBON β′2mp during sleep enhances long-term memory (Scaplen et al., 2020). Separately, local protein synthesis within MBON β′2mp, has been implicated in the consolidation of long-term memory (Wu et al., 2017). This makes MBON β′2mp an ideal model for understanding how sleep and memory signals might be integrated at a molecular level. It should be mentioned that MBON γ3β′1 reportedly acts as a sleep activator (Aso et al., 2014b) and local protein synthesis within this MBON is also important for the consolidation of long-term memory (Wu et al., 2017). Thus, MBON γ3β′one likely also plays a role in integrating sleep and memory signals through its reciprocal connections MBON β′2mp.

This provides a well-characterized anatomical framework to understand how opposing memories are acquired, consolidated, expressed and updated. Since the roles of these converging MBONs in naive and learned behaviors are state dependent (Grunwald Kadow, 2019; Lewis et al., 2015; Tsao et al., 2018), we hypothesize that MBON γ3β′1 and MBON β′2mp, both receiving convergent input from other MBONs, providing opportunities for feedforward networks to update information processing depending on the state of the animal.

Some of the feedback connections originally hypothesized to exist in the MB were between MBONs and DANs (Aso et al., 2014b; Ichinose et al., 2015). Our analysis revealed neurons postsynaptic to MBONs that are TH positive (Figure 3, Figure 8B). Recent studies that combined EM annotation and calcium imaging to identify specific MBON-DAN connections suggest extensive recurrent connectivity between MBONs and DANs, validating our findings (Felsenberg et al., 2018; Li et al., 2020; Otto et al., 2020). For example, previous studies using both GFP Reconstitution Across Synaptic Partners (GRASP) and EM annotation revealed that MBON α1 and DAN α1 are synaptically connected (Ichinose et al., 2015; Li et al., 2020). We similarly identified a few DAN neurons that innervate the horizontal MB lobes within the MBON α1 postsynaptic signal. A recent study showed that the 20 DANs that innervate the γ5 MB compartment are clustered into five different subtypes that innervate distinct anatomical regions within the γ5 compartment (Otto et al., 2020). According to this study, only one of the γ5 DANs receives direct recurrent feedback from γ5β′2a MBONs (Otto et al., 2020). Based on these recent anatomical characterizations, we believe that the TH+ neurons within the postsynaptic signal of γ5β′2a are the γ5 DANs.

The FSB is the largest substructure of the central complex, and it serves as a sensory-motor integration center (Pfeiffer and Homberg, 2014; Wolff et al., 2015). The FSB comprises nine horizontal layers (Wolff et al., 2015) that are innervated by large-field neurons (Hanesch et al., 1989). Previous work in blow flies (Phillips-Portillo and Strausfeld, 2012) and, later work in Drosophila (Aso et al., 2014a), predicted that the FSB was postsynaptic to output neurons of the MB. Our data confirm that the large-field, tangential neurons of the dorsal FSB are postsynaptic to the majority of MBONs. Although there exists some variation across brains (Figure 5—figure supplement 2), glutamatergic and GABAergic MBONs predominately project to FSB layers 4 and 5, whereas cholinergic MBONs mainly project to FSB layer 6. Connections between MBONs and FSB were consistent across different split-GAL4 lines that have overlapping expression patterns. Similar extensive direct connectivity between these MBONs and the dorsal FSB, especially layers 4 and 5, were found in the recently annotated EM hemibrain dataset (Li et al., 2020). Together, these observations suggest that the connectivity between the MB and FSB are structurally, and perhaps in some cases functionally, conserved across insect species.

How are FSB layers 4/5 and 6 functionally distinct? The dorsal FSB has a well-established role in modulating sleep and arousal (Berry et al., 2015; Donlea et al., 2011; Ueno et al., 2012), locomotor control (Strauss, 2002), courtship (Sakai and Kitamoto, 2006), and visual memory (Li et al., 2009; Liu et al., 2006; Wang et al., 2008). FSB layer 5 has been specifically implicated in processing information regarding elevation in a foraging- and rutabaga-dependent manner (Li et al., 2009). More recent studies have implicated the dorsal FSB in processing nociceptive information (Hu et al., 2018). FSB layer 6 plays a specific role in avoidance of a conditioned odor, whereas layers 4 and 5 respond to aversive stimuli and are responsible for innate, but not conditioned, avoidance (Hu et al., 2018). Moreover, recent connectome data suggest that differences exist in the postsynaptic connections of layers 4/5 and 6 as well. Overall, there is high degree of interconnectivity within the FSB (Clements et al., 2020). The predominate output of FSB layer 6 neurons are other FSB neurons. In fact, many FSB layer 6 neurons project exclusively to other FSB neurons (Clements et al., 2020). In contrast, FSB layer 4 neurons send direct projections to other brain structures - CRE, SMP, and LAL - in addition to projecting to other FSB neurons. The connections with the LAL position the FSB layer 4 to directly influence downstream motor output signals prior to executing behavior. Recent EM analysis also suggests that some FSB layer 6 neurons synapse back onto PAM DAN neurons (Li et al., 2020). This connectivity is in line with the associative role in conditioned nociception avoidance described for FSB layer 6 (Hu et al., 2018).

Interestingly, we found that the pattern of FSB postsynaptic targets of the MBONα1 is dissimilar to other glutamatergic MBONs. FSB layers 4/5 and 6 are not present in the MBON α1 postsynaptic signal. Instead, MBON α1 project to neurons that innervate the ventral and most dorsal aspect of the FSB. The ventral FSB is implicated in innate avoidance of electric shock (Hu et al., 2018), and more recent data suggest that its activity is tuned to airflow cues for orientation during flight (Currier et al., 2020). Artificial activation of MBON α1 does not result in significant avoidance behavior (Aso et al., 2014b). However, it has been implicated in the acquisition, consolidation, and expression of 24 hr long-term sucrose memory (Ichinose et al., 2015). It is possible that MBON α1 provides appetitive valence signals to the ventral FSB to guide goal-directed flight. Functionally validating the role of MBON α1 and its relationship with its putative downstream neurons is key to appreciating how learning signals can drive behavioral decisions.

More research is necessary to further understand the functional role of different FSB layers and how information is integrated across these layers. Based on the anatomical data, it is clear that although the MB and FSB can function in parallel during memory formation, they act as parts of a dynamic system to integrate information and adjust behavioral responses.

The LAL is an important premotor waystation for information traveling from the central complex to descending neurons innervating thoracic motor centers across insects (Chiang et al., 2011; Franconville et al., 2018; Hanesch et al., 1989; Wolff and Strausfeld, 2015a; Wolff and Strausfeld, 2015b). Accordingly, the LAL has been implicated in orientation to pheromones in the moth (Kanzaki et al., 1991a; Kanzaki et al., 1991b; Mishima and Kanzaki, 1999; Namiki et al., 2014; Namiki et al., 2018; Wada and Kanzaki, 2005), flight in the locust and dragonfly (Homberg, 1994; Olberg, 1986), locomotion in Drosophila (Bidaye et al., 2014) stimulus-directed steering in Drosophila, the cockroach, cricket, and moth (Harley and Ritzmann, 2010; Rayshubskiy et al., 2020; StaudacherY, 1998; Zorović and Hedwig, 2011) and in response to mechanosensory stimuli in the locust (Homberg, 1994). In the moth, recordings from neurons innervating the LAL have a characteristic ‘flip-flop’ firing property, which is thought to mediate walking commands (Kanzaki et al., 1991b; Kanzaki et al., 1994; Mishima and Kanzaki, 1998; Mishima and Kanzaki, 1999; Wada and Kanzaki, 2005). More recent work has suggested a functional organization whereby the neurons in the upper division of the LAL receive convergent input from the protocerebrum and neurons in the lower division generate locomotor command (Namiki et al., 2014; Rayshubskiy et al., 2020).

Our data show that the MB network converges with the protocerebrum input, thereby providing an opportunity for MBONs to indirectly influence descending motor outputs. We also demonstrate that two MBONs (γ3β′1 and γ2α′1) synapse on the same subset of LAL and FSB cells, revealing a convergent circuit that connects both structures. Further, in support of our anatomical observations, optogenetic activation of MBON γ2α′1 resulted in activation of both LAL and FSB layer four neurons. Given that MBON γ3β′1 is GABAergic, we did not perform the equivalent experiment for this neuron. Thus, understanding the functional consequences of these inhibitory connections will require further investigation. Interestingly, despite the fact that MBON γ3β′1 and γ2α′1 express different neurotransmitters and innervate different MB compartments, their manipulation has similar behavioral phenotypes: both promote sleep (Aso et al., 2014b; Sitaraman et al., 2015a; Sitaraman et al., 2015b), and artificial activation of either results in naive preference (Aso et al., 2014b). Further, activation of both MBON γ3β′1 and γ2α′1 together has an additive effect, which results in a significant increase in preference (Aso et al., 2014b).

The FSB and LAL have a well-established structural and functional connectivity. The LAL integrates information from the central complex, including the FSB, and provides a premotor signal to motor centers (Wolff and Strausfeld, 2015b). However, the behavioral significance of MBON γ3β′1 and γ2α′1 projections to both the FSB and LAL is less clear. Previous work demonstrated that activation of these MBONs while the flies explored an open arena did not significantly affect average speed or angular speed of individual flies (Aso et al., 2014b). By contrast, we found that inactivation of the putative downstream LAL neurons significantly increased overall activity of behaving flies in a social context and locomotor assay. Thus, the γ3β′1 and γ2α′1 MBONs may play a modulatory rather than required role in influencing behavioral response to an associated cue.

Recent work in Drosophila has demonstrated that the DANs that innervate MBON γ2α′1 regulate flight bout durations, and may provide a motivation signal via MBONs to the FSB and LAL to regulate motor activities (Sharma and Hasan, 2020). The LAL neurons receive multisensory input (Namiki and Kanzaki, 2016), and some LAL neurons make direct connections to descending neurons that control movement. Thus, this circuit organization enables integration of sensory signals with punishment or reward to direct the motion of the animal. In contrast, MBON connections with the FSB might play a role in providing context for flexible navigation, goal-directed actions, and memory-based navigation (Le Möel and Wystrach, 2020; Yue and Mangan, 2020).

If homology can be defined by shared expression of transcription factors and similar functional roles, the MB-FSB connection may be an appropriate model for understanding functional connections between the hippocampus and striatum (Strausfeld and Hirth, 2013a; Wolff and Strausfeld, 2016) and serve as an accessible model for understanding connectivity between more complex brain structures associated with memory. Further, given that the integrative relay role of the LAL is somewhat reminiscent of the vertebrate thalamus (Strausfeld and Hirth, 2013a), the complex connectivity between the MBONs, FSB, and LAL may also serve as an effective model for predicting and understanding functional connections between the hippocampus, striatum, and thalamus in the context of memory formation and action selection.

Insects exhibit a great variety of complex behaviors, and significant effort has been devoted to understand the neural circuits that underlie these behaviors. The genetically accessible Drosophila is a great model for studying the interplay between circuit architecture and behavior owing to their complex yet tractable brains. The MB circuits and their role in learning and memory are among the most studied circuits in Drosophila. Although, the majority of these studies have focused on olfactory memory, it is clear that the MB plays a much broader role in insect behavior. In Drosophila, the MB is important for courtship memory (McBride et al., 1999; Montague and Baker, 2016; Sitnik et al., 2003), taste aversive memory (Masek et al., 2015) as well as visual memory (Liu et al., 2006; Liu et al., 2016; Vogt et al., 2014). In cockroaches, the MB has a role in place memory (Mizunami et al., 1998) and recent data in two different species of ants implicate the MB in spatial navigation to learned locations using visual cues (Buehlmann et al., 2020; Kamhi et al., 2020). In mammals, the hippocampus is similarly required for multiple forms of associative memory, including spatial navigation using visual cues (Fenton et al., 2000a; Fenton et al., 2000b; Muller and Kubie, 1987; O'Keefe and Conway, 1978; Scaplen et al., 2014). Thus, cross-species similarity in circuit organization and function may exist between the mushroom body and the hippocampus (Wolff and Strausfeld, 2016). However, such anatomical and functional cross-species comparisons can also be made between the mushroom body and the cerebellum (Farris, 2011; Litwin-Kumar et al., 2017; Modi et al., 2020), suggesting that similar convergent-divergent architecture may be a general principle of structures that encode and update memories.

In this context, the implementation of trans-Tango to study the MB has high potential in the era of EM reconstruction of the Drosophila brain. Through examination of the circuit connectivity in several individuals, easily afforded by trans-Tango, the value of the EM reconstruction data could be augmented by overlaying on it potential nuanced differences between individuals. In addition, trans-Tango-mediated discoveries in the fly could help illuminate principles of circuit organization in other species. Further, due to the modular design of trans-Tango, it could be readily reconfigured for other types of studies beyond circuit tracing. For example, only minimal modifications are required for implementing a configuration of trans-Tango for identifying the molecular composition of the postsynaptic partners. This strategy could be used to examine the evidence that MBONs stratify the FSB through different classes of peptidergic neurons (Donlea et al., 2018; Kahsai et al., 2012; Kahsai and Winther, 2011; Nässel and Zandawala, 2019; Sareen et al., 2020). Confirmation of these observations would suggest that the MB plays a critical role in regulating modulatory systems of a midbrain region that shares structural and functional commonalities with the vertebrate basal ganglia (Strausfeld and Hirth, 2013a; Strausfeld and Hirth, 2013b). Finally, through combining it with new genome editing strategies, trans-Tango could become a useful tool for comparative anatomy in other insects. This would enable the study of synaptic connections in non-model organisms and lead to deeper understanding of biological diversity (Gantz and Akbari, 2018).

Understanding how memories are formed, stored, and retrieved necessitates knowledge of the underlying neural circuits. Our characterization of the architecture of the neural circuits connecting the MB with downstream central complex structures lays the anatomical foundation for understanding the function of this circuitry. Our studies may also provide insight into general circuitry principles for how information is processed to form memories and update them in more complex brains.

All Drosophila melanogaster lines were raised at 18°C on standard cornmeal-agar media with tegosept antifungal agent and in humidity-controlled chambers under 14/10 hr light/dark cycles. SS lines were previously made in the Rubin lab in collaboration with the Janelia FlyLight team and the Janelia Fly facility. For a list of fly lines used in the study, see the Key Resource Table.

Gibson Assembly was used to generate the plasmid UAS-CD2_QUAS-mtdTomato(3xHA). The DNA sequence encoding Rattus norvegicus CD2 (NP_036962.1) was codon optimized for Drosophila melanogaster and synthesized by Thermo Fisher Scientific, USA. This sequence was subsequently amplified using primers 5’-atcctttacttcaggcggccgcggctcgagaatcaaaATGCGCTGCAAGTTCCTG-3’ and 5’-agtaaggttccttcacaaagatcctctagaTTAGTTGGGTGGGGGCAG-3’ to obtain the insert fragment. To generate the vector fragment, the trans-Tango reporter plasmid (UAS-myrGFP_QUAS-mtdTomato(3xHA)) (Talay et al., 2017) was digested with XhoI and XbaI. Insert and vector fragments were ligated using HiFi DNA Assembly Kit (New England Biolabs, USA) following manufacturer’s instructions. The resultant plasmid was integrated at the su(Hw)attP8 site via PhiC31-mediated recombination.

Flies were dissected at 15–20 days post-eclosion using methods adapted from FlyLight Protocols (https://www.janelia.org/project-team/flylight/protocols). Flies were anesthetized with temperature, dewaxed in 70% ethanol, rinsed in Schneider’s Insect Medium (S2) and dissected on a Sylgard pad with cold S2. Within 20 min of dissection, collected brains were transferred to 2% paraformaldehyde (PFA) in S2 and incubated for 55 min at room temperature. After fixation, brains were rinsed with phosphate buffered saline with 0.5% Triton X-100 (PBT) for 15 min at room temperature. Washes were repeated four times before storing the brains overnight in 0.5% PBT at 4°C. For chemical tagging in brp-SNAP+ brains, PBT was removed and SNAP substrate diluted in PBT (SNAP-Surface649, NEB S9159S; 1:1000) added. Brains were incubated for 1 hr at room temperature and rinsed with PBT (3 times for 10 min). Brains were then blocked in 5% GS (Goat Serum) diluted in PBT for 90 min at room temperature. Brains were then incubated in primary antibodies diluted in 5% GS/PBT for 4 hr at room temperate and then at 4°C for two overnights. After primary antibody incubation, brains were washed four times for 10 min with 0.5% PBT before incubating in secondary antibodies diluted in 5% GS/PBT at 4°C for two overnights. Samples were then rinsed and washed four times for 15 min in 0.5% PBT at room temperature and prepared for DPX mounting. Briefly, brains were fixed a second time in 4% PFA in PBS for 4 hr at room temperature and then washed four times in PBT for 15 min at room temperature. Brains were rinsed for 10 min in PBS, placed on PLL-dipped cover glass, and dehydrated in successive baths of ethanol for 10 min each. Brains were then soaked three times in xylene for 5 min each and mounted using DPX.

MBON split-GAL4 ‘C’ lines which have the DNA-binding domain (in attP2) and activation domain (in VK00027) recombined on the 3rd chromosome were crossed to newly generated trans-Tango reporter flies where the 10xUAS-myrGFP was replaced with 10xUAS-CD2, and 13xLexAOp-mCD8::GFP was inserted into attP2. This enabled the visualization of the starter MBONs, the postsynaptic trans-Tango signal, and the LexA lines simultaneously.

Confocal images were obtained using a Zeiss, LSM800 (Brown University) and LSM710 (Janelia Research Campus) with ZEN software (Zeiss, version 2.1) with auto Z brightness correction to generate a homogeneous signal and were formatted using Fiji software (http://fiji.sc). Whole brains were scanned using a 40x objective in four overlapping tiles and then stitched together in the ZEN software.

TH+ cells, and cells with overlapping TH and trans-Tango signal were counted by blinded experimenter using the Cell Counter plugin in FIJI (https://imagej.net/Cell_Counter). We counted the total number of TH+ cells that co-localized with trans-Tango labeled cells in each hemibrain starting at the most anterior surface of the brain and continued to count TH+ cells until we reached the protocerebral anterior lateral (PAL) cluster which were identified by their cell body size. We did not identify any co-localized cells within or posterior to the PAL cluster.

Images were prepared for publication in FIJI and Adobe Illustrator with no external manipulation aside from cropping to demonstrate higher resolution. All figures were generated using Adobe Illustrator CC.

Brains were registered as previously described (Aso et al., 2014a). Postsynaptic connections of registered brains were segmented in VVD Viewer (https://github.com/takashi310/VVD_Viewer; Wan et al., 2009; Wan et al., 2012) and saved as .nrrd files. Segmented files of postsynaptic signal for each MBON were multiplied by 34 binary masks of each central brain region in a custom written Matlab program to calculate the distribution of postsynaptic signal across brain regions. Signal within each brain was normalized by calculating a Z-score, or the number of standard deviations above or below the mean signal, for each brain regions. Heatmaps were generated in RStudio.

All functional imaging experiments were performed ex-vivo from brains of 1- to 4-day-old male or female brains on an Ultima two-photon laser scanning microscope (Bruker Nanosystems) equipped with galvanometers driving a Chameleon Ultra II Ti-Sapphire laser. Images were acquired with an Olympus 60x, 0.9 numerical aperture objective at 512 × 512 pixel resolution.

Flies were placed on food containing 400 μM all trans-retinal for 18–36 hr prior to dissection. Brains were dissected in saline (108 mM NaCl, 5 mM KCL, 2 mM CaCl2, 8.2 mM MgCl2, 4 mM NAHCO3, 1 mM NaH2PO4, 5 mM trehalose, 10 mM sucrose, 5 mM HEPES, pH 7.5 with osmolarity adjusted to 275 mOsm), briefly (45 s) treated with collagenase (Sigma #C0130) at 2 mg/mL in saline, washed, and then pinned with fine tungsten wires in a thin Sylgard sheet (World Precision Instruments) in a 35 mm petri dish (Falcon) filled with saline. MBONs were stimulated with 400-500ms of 627 nm LED. For recordings in the LAL (VT018476 and VT055139) ROI were positioned over SMP. For recordings in the FSB (476H09) ROIs were positioned over SMP or FSB.

All image processings were done using FIJI/ImageJ (NIH). Further analysis was performed using custom scripts in ImageJ, Microsoft Excel, and RStudio. Normalized time series of GCaMP fluorescence were aligned to the time point when the opto-stimulus was applied for each replicate.

Locomotor activity was evaluated in a 37 mm diameter circular open field area as described previously (Scaplen et al., 2019). Briefly, for thermogenetic inactivation, 10 flies were placed into arena chambers and placed in a 30°C incubator for 20 min prior to testing. The arena was then transferred to a preheated (30°C) light sealed box and connected to a humidified air delivery system. Flies were given an additional 15 min to acclimate to the box before recordings began. Group activity was recorded (33 frames/s) for 20 min. Recorded .avi files of fly activity were processed by FFMPEG and saved as .mp4. Individual flies were tracked using Caltech Flytracker (Eyjolfsdottir et al., 2014) to obtain output features such as position, velocity, and angular velocity. Feature based activity was averaged across within each genotype and plots were generated in RStudio.