Sensory neurons couple arousal and foraging decisions in C. elegans

Foraging animals optimize feeding decisions by adjusting both common and rare behavioral patterns. Here, we characterize the relationship between an animal’s arousal state and a rare decision to leave a patch of bacterial food. Using long-term tracking and behavioral state classification, we find that food leaving decisions in C. elegans are coupled to arousal states across multiple timescales. Leaving emerges probabilistically over minutes from the high arousal roaming state, but is suppressed during the low arousal dwelling state. Immediately before leaving, animals have a brief acceleration in speed that appears as a characteristic signature of this behavioral motif. Neuromodulatory mutants and optogenetic manipulations that increase roaming have a coupled increase in leaving rates, and similarly acute manipulations that inhibit feeding induce both roaming and leaving. By contrast, inactivating a set of chemosensory neurons that depend on the cGMP-gated transduction channel TAX-4 uncouples roaming and leaving dynamics. In addition, tax-4-expressing sensory neurons promote lawn-leaving behaviors that are elicited by feeding inhibition. Our results indicate that sensory neurons responsive to both internal and external cues play an integrative role in arousal and foraging decisions.


21
Persistent behavioral states subdivide continuous behavior into discrete modules that accomplish 22 adaptive goals (Tinbergen, 1951). An example from behavioral ecology is foraging behavior, which is 23 composed of locomotion patterns that unfold across short and long timescales. For example, the brief 24 darting maneuvers that drive prey capture in hunting zebrafish are embedded within persistent 25 locomotory arousal states that last for minutes (Marques et al., 2020). In recent years, classical 26 ethological studies of behavioral states have been supplemented with machine vision, clustering, and 27 classification algorithms that are well-suited to quantitative and systematic analysis (Berman et al., 2014; 28 Schwarz et al., 2015;Wiltschko et al., 2015). A current challenge is to identify circuit mechanisms that 29 couple long-term behavioral states to short-term motor actions in foraging and other goal-directed 30 behaviors.

31
Foraging behaviors and the neural mechanisms that generate them have been a fruitful subject of 32 study in the nematode Caenorhabditis elegans. While exploring a bacterial lawn, which corresponds to a

50
The genes that regulate lawn-leaving behavior overlap with those that regulate roaming and 51 dwelling states, although these behaviors have largely been studied separately. For example, roaming, 52 dwelling, and leaving are all strongly influenced by sensory neurons, particularly those that express the

57
Here, we characterize the behavioral patterns and circuit mechanisms that accompany lawn 58 leaving decisions on short and long timescales. Using high-resolution imaging to monitor the behaviors of 59 many individual animals, we find that leaving is a discrete behavioral motif that is generated 60 probabilistically during the high-arousal roaming state. A signature of leaving behavior is a rapid 61 acceleration in speed immediately before the leaving event. Roaming states evoked by neuromodulatory  9 Lawn leaving behavior is associated with high arousal states on short and long timescales

135
To gain further insight into the behavioral states preceding lawn leaving, we coarse-grained our 136 behavioral measurements into 10 second intervals ("bins") and expanded the time axis to allow analysis 137 across different durations. On average, a slow rise in speed began two minutes before lawn leaving, with 138 a rapid acceleration in the last minute before leaving ( Fig. 2A). We analyzed these behaviors in the  Table 2). Five minutes before leaving, 34% of animals were roaming, a 146 fraction that steadily increased such that 80% of animals were roaming immediately before leaving ( Fig.   147 2C, Supplementary Table 2). By contrast, the fraction of animals roaming before head poke-reversals 148 increased only slightly before the event (Fig. 2-S3C). Within roaming states, the average speed was 149 constant until the rapid acceleration ~30 seconds prior to lawn leaving (Fig. 2D). To ask if the brief 150 acceleration accounts for the apparent increase in roaming before lawn leaving, we compared the 151 duration of roaming states that preceded lawn leaving to those that did not. We found that roaming states 152 directly before lawn leaving were slightly longer than other roaming state, arguing that acceleration events 153 alone do not drive the correlation between roaming and lawn leaving (Fig. 2-S1E).

154
The roaming-dwelling HMM was initially developed on uniform bacterial lawns (Flavell et al., 155 2013), which elicit more roaming than the small lawns used here (Fig. 2-S1F

156
Expanding this model to include posture information identified eight subclasses of dwelling states 157 (Cermak et al., 2020). To explore alternative analysis methods, we generated an HMM behavioral state  Table 2), but five minutes before leaving, 26% of animals were in state 3, which 165 ramped to 82% immediately before leaving (Fig. 2F, Supplementary Table 2). Like roaming, state 3 166 increased only slightly before head poke reversals (Fig. 2-S3E). Animals across all behavioral states -167 roaming, dwelling, and the four AR-HMM states -were similarly distributed across the bacterial lawn, with 168 strong enrichment near the lawn boundary ( Fig. 2-S4A-D).

169
Both the two-state model and the four-state model show that arousal states increase before lawn 170 leaving over at least two phases: an enrichment of a high arousal state 3-5 minutes before leaving, 171 followed by a rapid acceleration in the last 30 seconds. Because the widely used two-state roaming-172 dwelling model captured the key features of lawn-leaving as well as these alternative methods, we 173 applied that analysis to subsequent experiments.

267
Dark lines represent the mean and shaded region represents the standard error.

268
In boxplots, median is highlighted, and interquartile range is indicated by shaded area.

269
See Supplementary

280
(D) Same as (C) but plotted as a cumulative distribution.

281
Distributions are computed across wild type animals on small lawns (n = 1586).

282
In A-B, **** p < 10 -4 . In C-D, each AR-HMM state is significantly different from all others (a-d), although 283 effect sizes are small.

301
While roaming and head pokes began immediately upon feeding inhibition, leaving events did not 302 occur immediately but instead accumulated throughout the 10-minute stimulation interval (Fig. 3I). In each 303 case, leaving was preceded by a 30 second acceleration in speed (Fig. 3J). Thus, feeding inhibition elicits 304 both roaming and lawn-leaving behaviors, and lawn-leaving occur probabilistically during roaming states.

335
In time-averages (F,J), dark line represents the mean and shaded region represents the standard error.

336
Violin plots show median and interquartile range.

(A-B)
Aztreonam affects roaming and leaving by affecting bacterial growth. In "pre-add" conditions, the 346 drug is added to liquid cultures and agar plates during bacterial growth. In "post-add" conditions, the drug 347 is added after bacterial growth just before testing behavior.

370
Statistics by Mann-Whitney U test.

372
Violin plots show median and interquartile range.

373
In (D,E,G), each dot pair connected by a line represents data from a single animal. In (C), each dot pair 374 connected by a line represents data preceding a single lawn leaving event.

378
To further examine the relationship between arousal states and lawn leaving, we examined mutants with 379 known alterations in roaming and dwelling. Animals deficient in serotonin (tph-1), dopamine (cat-2), or the 380 neuropeptide receptor NPR-1 (npr-1) roam at a higher rate than wild type ( . We found that each of these 382 mutants showed increased lawn leaving compared to wild-type controls, strengthening the observed 383 correlation between roaming and leaving rates (Fig 4A-C, E-G, Table 2). In addition, the fraction of 384 animals roaming increased over several minutes before leaving events, as was observed in wild type

391
Roaming rates increased similarly during the 3 minutes prior to lawn leaving in wild type and pdfr-1 392 animals, suggesting that the coupling of roaming and leaving does not require PDFR-1 signaling (Fig. 4L).

393
Although their basal locomotion speed is lower (

421
In time-averages, dark line represents the mean and shaded region represents the standard error.

453 454
Acute circuit manipulation drives deterministic roaming and probabilistic leaving 455 As a complement to the neuromodulatory mutants, we employed a circuit-based approach to manipulate 456 arousal levels and examine effects on lawn leaving. Roaming is strongly stimulated by pdfr-1; we defined 457 sites of pdfr-1 expression that stimulate roaming using an intersectional Cre-Lox system that restores       Flavell, et. al (2013) and this work (C)                 (Fig. 6A, Table 1). However,

556
we found that tax-4 mutant animals continued to leave lawns -indeed, they left at slightly higher rates 557 than wild-type animals (Fig. 6B, Table 2). Moreover, the temporal relationship between roaming and 558 leaving was altered in tax-4 mutants, which typically roamed for only ~1 minute prior to lawn leaving ( Fig.   559 6C, Fig. 6-S1A-B). These results indicate that loss of tax-4 disrupted the characteristic arousal dynamics 560 associated with leaving.

561
Many of the 15 classes of sensory neurons that express tax-4 have been implicated in roaming or 562 leaving behaviors (Table 1). We rescued tax-4 separately in AWC, which senses food odors; ASK, which 563 senses amino acids and pheromones; ASJ and ASI, which sense pheromones, food, and toxins; and

568
6A-C, Fig 6-S1), and tax-4 rescue in the ASK neurons, which paradoxically suppressed leaving to a level 569 below that of either wild type or tax-4 animals ( Fig. 6-S2). Because ASJ and ASK showed opposite 570 effects in these experiments, we also examined strains in which both neurons were rescued. Combined 571 ASJ and ASK rescue normalized roaming and leaving compared to ASK rescue alone, albeit not to fully 572 wild-type levels ( Fig. 6-S2). While we have not tested all neurons and combinations, these results 573 suggest that multiple tax-4-expressing sensory neurons have roles in arousal-related behaviors and 574 highlight ASJ as a regulator of roaming and leaving dynamics.

575
Next, we asked how tax-4 mutants responded to acute inhibition of feeding. As in wild-type 576 animals, optogenetic inhibition of pharyngeal pumping resulted in an immediate and strong increase in 577 roaming in tax-4 mutants (Fig. 6D, Fig. 6-S3A). However, feeding inhibition in tax-4 mutants increased 41 leaving only slightly, unlike in the wild-type ( Figure 6E, Table 2). Both of these effects were rescued by 579 expressing tax-4 in ASJ neurons (Fig. 6C-E, Fig. 6-S3B-D). Similarly, tax-4 animals on inedible food 580 roamed at the same high rate as wild type animals but produced significantly fewer lawn leaving events 581 ( Fig. 6-S4); rescuing tax-4 in ASJ restored lawn leaving to wild-type levels. Thus tax-4 sensory mutants 582 uncouple leaving behavior from its normal context in multiple ways: they can leave edible food lawns 583 without an extended roaming state, and they are less likely to leave when feeding is inhibited, even while 584 they are roaming.

720
Our results identify two distinct biological mechanisms that regulate lawn leaving. First, leaving is coupled 721 to arousal state. Leaving rates are 20-fold higher in roaming animals than in dwelling animals, and the 722 leaving rates of most arousal mutants are largely explained by the fact that they spend more time roaming

723
( Table 2). The second mechanism is controlled by tax-4 sensory neurons, which shape the behavioral

730
Accordingly, many sensory neurons affect roaming, dwelling, or leaving (Table 1)  Others have shown that chronic feeding deprivation across hours drives lawn leaving through the 737 action of tax-4 in several sensory neurons (Olofsson, 2014). An unexpected result obtained here was that 738 tax-4 sensory neurons were required to drive the high leaving rates after acute inhibition of feeding.

739
Together, these results suggest that the sensory neurons are sites at which internal feeding cues,    Leaving events while roaming are increased by feeding inhibition in wild type but not in tax-4 mutants 58 773

823
Small lawn foraging assay

824
For all assays, E. coli OP50 was grown overnight in a shaking LB liquid culture from a single colony at 37ºC. On 825 the morning of the assay, 400 μL of saturated liquid culture was diluted into 5 mL of LB and allowed to grow to 826 OD1 at 37 ºC (~1.5 hours), as measured by spectrophotometer. The liquid culture was then spun down and 827 resuspended in M9 buffer (3 g KH2PO4, 6 g Na2HPO4, 5 g NaCl, 1 ml 1 M MgSO4, H2O to 1 liter) then 828 concentrated to a density of OD2 (and OD1 or OD4 in Figure 3A

832
50 μL of bacterial resuspension was seeded onto a separate NGM agar plate to be used as a food density 833 acclimation plate. Lawns were grown at 20-22ºC for 2 hours before the assay. Adult hermaphrodites picked as

834
L4s 16-20 hours before the assay were then transferred to acclimation plates. After 45-90 minutes, animals were 835 transferred to an unseeded NGM plate, cleaned of E. coli, and transferred singly into each well of the assay 836 plates, where they were placed on bacteria-free agar and allowed to find the small food lawn on their own.

837
Animals of the same genotype were grouped on the same 6-well plates and each plate was recorded by a single 838 camera. We used 12 cameras, enabling simultaneous recording of up to 72 individual animals at a time.

839
Temperature and relative humidity within the behavioral recording apparatus were continuously monitored during 840 recordings to ensure that environmental conditions were consistent across filming locations. As a further precaution, the filming locations of each genotype and wild type controls within the recording apparatus were randomized across batches of experiments and days to average out behavioral influences deriving from nonuniform local environmental conditions. Assays were recorded for 1 hour at 3 frames per second using 12 8.8 MP data were retained for analysis to minimize the effects from manipulating animals prior to recordings. Data from power and a 5% alpha, yielding a minimum value of n=18 animals per group for a two-sample unmatched