Tailup expression in Drosophila larval and adult cardiac valve cells

In Drosophila larvae, the direction of blood flow within the heart tube, as well as the diastolic filling of the posterior heart chamber, is regulated by a single cardiac valve. This valve is sufficient to close the heart tube at the junction of the ventricle and the aorta and is formed by only two cells; both are integral parts of the heart tube. The valve cells regulate hemolymph flow by oscillating between a spherical and a flattened cell shape during heartbeats. At the spherical stage, the opposing valve cells close the heart lumen. The dynamic cell shape changes of valve cells are supported by a dense, criss‐cross orientation of myofibrils and the presence of the valvosomal compartment, a large intracellular cavity. Both structures are essential for the valve cells' function. In a screen for factors specifically expressed in cardiac valve cells, we identified the transcription factor Tailup. Knockdown of tailup causes abnormal orientation and differentiation of cardiac muscle fibers in the larval aorta and inhibits the formation of the ventral longitudinal muscle layer located underneath the heart tube in the adult fly and affects myofibrillar orientation of valve cells. Furthermore, we have identified regulatory sequences of tup that control the expression of tailup in the larval and adult valve cells.


Summary
In Drosophila larvae, the direction of blood flow within the heart tube, as well as the diastolic filling of the posterior heart chamber, is regulated by a single cardiac valve.
This valve is sufficient to close the heart tube at the junction of the ventricle and the aorta and is formed by only two cells; both are integral parts of the heart tube. The valve cells regulate hemolymph flow by oscillating between a spherical and a flattened cell shape during heartbeats. At the spherical stage, the opposing valve cells close the heart lumen. The dynamic cell shape changes of valve cells are supported by a dense, criss-cross orientation of myofibrils and the presence of the valvosomal compartment, a large intracellular cavity. Both structures are essential for the valve cells' function. In a screen for factors specifically expressed in cardiac valve cells, we identified the transcription factor Tailup. Knockdown of tailup causes abnormal orientation and differentiation of cardiac muscle fibers in the larval aorta and inhibits the formation of the ventral longitudinal muscle layer located underneath the heart tube in the adult fly and affects myofibrillar orientation of valve cells. Furthermore, we have identified regulatory sequences of tup that control the expression of tailup in the larval and adult valve cells.

K E Y W O R D S
cardiogenesis, dorsal vessel, drosophila, heart tube, metamorphosis, tailup, valve cells

RESULTS AND DISCUSSION
The larval Drosophila heart constitutes a tubular, two-chambered linear organ built by contractile cardiomyocytes that possess an internal structure comprising repetitive actin-myosin sarcomeres, similar to skeletal muscles. The anterior aorta and the posterior heart proper, or heart chamber, are physically separated by a single pair of cardiomyocytes, differentiated into highly specialized valve cells during embryogenesis (Lammers et al., 2017). Valve cells harbor large vesicular membrane compartments-valvosomes-which are crucial for cell shape dynamics and cell stability when the heart is beating. Recently, we have shown that valvosomes originate from redirected endosomal pathways and are maintained by fusion with recycling endosomes (REs) (Meyer et al., 2022).
The intracardiac valves are critical for regulating flow directionality within the heart tube. They ensure the efficient streaming of hemolymph from the heart chamber toward the anterior opening of the aorta (Lammers et al., 2017). The larval heart harbors one valve, whereas the heart of the adult fly has three. The two additional valves differentiate in the course of hormonally regulated partial remodeling of the larval heart during metamorphosis, resulting in a linear heart tube with four separated ventricles (Monier, Astier, Sémériva, & Perrin, 2005). The single larval valve is maintained into adulthood. The two additional valves differentiate from specific cardiomyocytes, giving rise to three valves in total in the imago (Lammers et al., 2017;Lehmacher, Abeln, & Paululat, 2012;Rotstein & Paululat, 2016;Tang, Yuan, Bodmer, Wu, & Ocorr, 2014;Zeitouni et al., 2007). The myofilaments of valve cells exhibit a cross-helical orientation, whereas those of all other cardiomyocytes are predominantly circular. Contraction of myofibres turns the valve cells from spherical into a flattened shape, corresponding to the open or closed state of the valve, respectively (Lammers et al., 2017). The circular orientation of myofibres in other cardiomyocytes provides a posterior-to-anterior continuous wave of contraction of the heart tube and, thus, continuous motion of hemolymph.
By screening for transgenic GFP-and Gal4-reporter lines, showing expression in heart cells and particularly the valves, we identified the 76E11-Gal4 line, which harbors regulatory elements of the tailup (tup) gene. We found 76E11 to be initially active in all cardiomyoblasts of the embryo and early larvae (Morin-Poulard et al., 2022). The tup gene encodes a LIM domain transcription factor that contains a homeo-and a LIM domain DNA-binding motif and has been previously shown to play a role in cardiac cell specification and differentiation. However, the broad cardiac enhancer activity of 76E11 seen in the early larvae diminishes during the second larval stage and becomes restricted to the single pair of valve cells in third-instar larvae (Figure 1a-e) and the three pairs of valve cells in adult flies (Figure 1f,g). Expression is maintained throughout the lifetime without diminishing upon aging. In addition to expression in cardiac tissue, 76E11 is active in adult alary muscles. A closer analysis of 76E11 expression revealed reporter gene expression also in the anterior cardiomyocytes immediately adjacent to the valve cells, although often at a lower level of expression ( Figure 1g). The 76E11-Gal4 reporter line harbors a regulatory element of the tup gene and was generated in a larger screen for the neurogenic activity of enhancer sequences (Jenett et al., 2012). The 3 kb genomic regulatory element present in 76E11 is located 22.5 kb upstream of the transcriptional start site of tup ( Figure 1j). In addition to expression in the cardiac valves, we also see reporter activity of 76E11-Gal4 in yet-unidentified cells of the brain and the ventral nerve cord.
Additional tup-reporter lines have been generated and tested for their expression during embryogenesis (Boukhatmi et al., 2014;Tao, F I G U R E 1 Enhancer activity of 76E11-Gal4 and tup-F4-GFP in the heart. (a-e) The 76E11-Gal4 driver line shows strong reporter gene expression exclusively in the valve cells of third-instar larvae (a, asterisks label valves). Valve cell (vc) characteristic valvosomes (c, asterisks) are visualized by expression of a membrane-bound GFP reporter (arrows in d,e); (b, Maximum projection; c-e, selected single slices). (f and g) In adult flies, vc specificity vanishes, and 76E11-Gal4 is capable of driving expression of the reporter gene in valve cells (asterisks) and adjacent cardiomyocytes (cm), as well as alary muscles (am). In contrast to the handC-GFP reporter, 76E11-Gal4 is not active in all cardiomyocytes (g). (h and i) tup-F4 is sufficient for expression in all cm including valve cells (asterisks) am of third-instar larvae and adult flies. (j) Mapping of regulatory elements upstream of the tup genomic region. pn, pericardial nephrocytes Wang, Tokusumi, Gajewski, & Schulz, 2007). One of these lines is tupF4, which contains a 1.5 kb regulatory sequence, located 11.5 kb upstream of the transcriptional start site of tup (Figure 1j), and drives expression in all cardioblasts, pericardial nephrocytes, lymph gland cells, and alary muscles of the embryo (Boukhatmi et al., 2014;Tao et al., 2007). We tested tupF4 for later expression at larval and adult stages and found reporter gene activity in third-instar larvae in all car- To analyze the developmental origin of the 76E11-Gal4 expressing valve cells in more detail, we utilized the G-trace system, a Gal4-based technique that allows lineage tracing of fluorescent reporter gene-labeled cells (Evans et al., 2009). The advantage of this method is the Gal4-dependent initiation of reporter expression, whereas the maintenance of reporter expression is not dependent on this initiation. Cells currently expressing Gal4 are marked by RFP, whereas cells that have been expressing Gal4 in the past are expressing GFP constantly. We found that 76E11 drives the expression of RFP only in the valve cells of third-instar larva, whereas the lineage expression is more expanded (Figure 3a to that in third-instar larvae. The 76E11 driver was active in all heart cells including cardiomyocytes and pericardial nephrocytes, as well as in the alary muscles and muscles of the heart tube. In summary, our experiment demonstrates further that the 76E11 enhancer element is initially broadly expressed in the cardiac cell lineage but later in development becomes restricted to valve cells and alary muscles ( Figure 3h).
The expression of Tup in all cardiomyocytes from embryogenesis to adulthood and the presence of specific enhancers that provide high expression in late cardiac valve cells raises the question of whether the transcription factor Tup is required for the specification, differentiation, or function of these cells. Tup homozygous amorph mutants have previously been shown to die during late embryogenesis. The mutant embryos exhibited various developmental defects, including cardiac defects; for example, the number of cardioblasts is altered, and heart closure is impaired (Mann, Bodmer, & Pandur, 2009). Tup interacts genetically with Mef2, Tinman and Pannier, three of the key regulators of cardiogenesis (Mann et al., 2009). In addition, Tup has been shown to play an important role in both the developing and mature nervous systems (Santiago & Bashaw, 2017;Thor & Thomas, 1997;Wolfram, Southall, Brand, & Baines, 2012). In the mesoderm, Tup is required not only for the differentiation of the cardiac cells but also for the differentiation of the heart-associated alary muscles and the recently identified thoracic alary-related muscles (TARMs) (Bataillé et al., 2020;Boukhatmi et al., 2014), as well as the cardiac outflow tract (Zmojdzian & Jagla, 2013). Furthermore, Tup plays an important role in hematopoietic organ formation (Tao et al., 2007). Because homozygous tup mutants die during embryonic or early larval development and, thus, do not allow conclusions to be drawn of possible late functions in the heart or adjacent tissues, we analyzed RNAi-mediated knockdown animals (Figure 4-6). First, we down-regulated tup using the 76E11-Gal4 driver, characterized in this article ( Figure 1). As an independent driver line for tup down-regulation, we used mef2-Gal4, which is active in all cardiomyocytes including valve cells, in somatic, visceral, and in alary muscles. We stained mutant and wild-type larval hearts for alpha-Spectrin to visualize the F I G U R E 2 Expression of Tup in heart tissue. Dissected thirdinstar wandering larvae and dissected 1-week-old adult specimens were stained with anti-Tup antibodies (red channel). Sarcomere structure was visualized by F-actin staining (green channel). (a) Tup is detectable in larvae in the nuclei of all cardiomyocytes (cm), including valve cells (asterisks). (b) In the adult specimen, all cardiomyocytes, including valve cells, alary muscles, somatic muscles (arrow), and additional tissue express Tup. (c-g) Cardiomyocytes (c), valve cells of the larval heart (d-f), and all cardiomyocytes and valves of the adult heart (g) were stained with anti-Tup antibodies (red channel) and anti-GFP antibodies to visualize the activity of the tupF4-GFP reporter (green channel). cm, cardiomyocyte; lg, lymph gland; os, ostia. Dashed lines in (c) and (g) label the cardiac tube unique histology of the valve cells (Figure 4a-e) and found that valve cells differentiate properly upon tup knockdown. The cell size is unchanged in animals in which Tup is down-regulated with either 76E11-Gal4 or mef2-Gal4 drivers. The valvosomes, which are intracellular cavities responsible for cell shape, are also present in normal numbers (Meyer et al., 2022). Furthermore, valve cells are built at their defined position in the heart (Figure 4a-g). However, when we stained specimens for F-actin to visualize muscle organization, we noticed that the myofibres in cardiomyocytes completely or partially lost their typical circular arrangement ( Figure 5). This phenotype is most prominent when we combined mef2-Gal4 as the driver line and the more efficient of the two tup RNAi lines, tup RNAi #2. In these animals, all myofibrils in cardiomyocytes are oriented fully longitudinally rather than circularly (Figure 5e). The same accounts for the intracardiac valves, which display exclusively longitudinal-orientated myofibrils (Figure 5e 0 ), rather than a criss-cross network organization (Figure 5a 0 ,c 0 ,g 0 ). However, myofibrils are still functional and able to contract, but they contract exclusively in an anterior-posterior direction (Movie S1). A similar phenotype has been noticed upon ectopic expression of Abdominal-A in the larval aorta (Monier et al., 2005) and also for tinman mutants (Zaffran et al., 2006), supporting the possibility that Tup may play a role as a regulator of Tinman (Tin) expression. Furthermore, similar changes in myofibril orientation have been observed in animals, mutant for the ECM protein Pericardin, in which the alary muscles detach from the heart (Drechsler et al., 2013).
Indeed, knockdown of tup, using mef2-Gal4, results in a strong reduction or complete absence of alary muscles in third-instar larvae when crossed to the stronger tup RNAi #2 line ( Figure S1). Hence, we conclude that Tup is crucial for the correct development of alary muscles.
However, an additional impact on myofiber orientation in adult cardiomyocytes is provided by the so-called ventral longitudinal musculature (VLM). VLMs are situated in parallel rows underneath the heart tube in an anterior-posterior orientation (Lehmacher et al., 2012;Schaub, März, Reim, & Frasch, 2015). This sheet of underlying longitudinal musculature is supposed to support the heart tube structurally. VLMs originate during metamorphosis from alary muscles as part of a reprogramming process. Transformation of alary muscles is tightly controlled by the transcription factor Org-1 (Boukhatmi et al., 2014;Schaub et al., 2015;Schaub, Rose, & Frasch, 2019). Org-1 mutants fail to form alary muscles and therefore also VLMs, which results in malformations of the heart tube including myofiber misorientation in cardiomyocytes (Schaub et al., 2015). Absence of the ventral longitudinal muscle (VLM) is also induced upon depletion of tup specifically in larval alary muscles using org1-Gal4 as a driver line (Schaub et al., 2015). Importantly, the org1 enhancer is active specifically in the alary muscle lineage but lacks activity in cardiac tissues.

F I G U R E 3
Characterization of real-time and lineage expression of the tup 76E11-Gal4 driver using the G-trace system. (a-d) The heart of third-instar larva show 76E11 driver activity exclusively in the valve cells (asterisks). Lineage-specific expression marked by GFP supports the former enhancer activity in lymph glands (lg), cardiomyocytes (cm), alary muscles (am), and pericardial nephrocytes (pn). (e-g) Adult fly heart. Realtime expression of 76E11-driven Gal4 is observed in all three pairs of valve cells (asterisks) and the adjacent cardiomyocytes (cm). In addition, the driver is active in the alary muscles (am). Earlier driver activity is found in all cardiac cells, alary muscles, and muscles of the heart tube. (h) Schematic summary of the 76E11-Gal4 driver real-time and lineage expression in third-instar larvae and adult flies.
Here, we find, that animals in which we downregulated tup by mef2-Gal4, display also a lack of the VLM (Figure 6d-e 0 ). This result therefore supports previous observations. However, when we used the 76E11-Gal4-driver to knockdown tup, this had no impact on VLM formation (Figure 6a-b 0 ) and also no impact on myofibril orientation in cardiomyocytes. Therefore, we conclude that correct myofibril orientation in cardiomyocytes depends on external guidance provided by the alary and/or VLMs. This conclusion is further supported by our observation that the development of alary muscles is already impaired in larval stages upon knockdown of tup. To support our results by an additional experiment, we downregulated Tup using hand-Gal4 and analyzed myofibril organization in cardiomyocytes (Figure 5f, g). Importantly, the hand enhancer used in this Gal4-driver line is active in cardiomyocytes (and pericardial nephrocytes) but not in alary muscles, ventral longitudinal or any other somatic muscle tissue. Therefore, by comparing the analyzed driver lines, we were able to decide whether Tup in cardiomyocytes is directly necessary for myofibril orientation or whether myofibril orientation is solely determined by external influences. We found that hand-driven knockdown of tup had no effect on myofibril orientation of cardiomyocytes (Figure 5f, g). In addition, VLM formation is not affected by hand-driven tup knockdown (Figure 6f-g 0 ). Our experiments show that the orientation of myofilaments in cardiomyocytes is not controlled by the cardiomyocytes themselves but is governed by external structural cues, possibly by the architecture of alary muscles or the VLMs (Figure 6). It might be that the VLMs provide anchor points for mechanical coupling with the heart tube. This supposition will be investigated in future work. is down-regulated in alary muscles, leading to their disappearance, there will be defects in the myofibril orientation in larval cardiomyocytes and differentiation deficits of the adult heart during metamorphosis.

Fly stocks
The following fly stocks were used in this study: handC-GFP ( F I G U R E 5 Muscle fiber organization and the morphology of the larval aorta are severely affected upon tup knockdown. (a-g) Show the heart tube of third-instar larvae, (a 0 ,c 0 ,e 0 ,g 0 ) show valve cells of larva, position indicated by dashed yellow lines. (b,c) downregulation of tup, two RNAi-constructs, with 76E11-Gal4. (d,e) Downregulation of tup, two RNAi-constructs, with mef2-Gal4. (f,g) Downregulation of tup, two RNAi-constructs, with handC-Gal4. In general, the tup RNAi-line #1 is less efficient than the tup RNAi-line #2. The strongest phenotype is observed in larval hearts when the tup RNA-line #2 is combined with mef2-Gal4 as the driver (e). This combination leads to a complete re-orientation of myofibres in cardiomyocytes and intracardiac valves (e 0 ). Circularly oriented muscles are no longer visible, and alary muscles (am) are absent in third-instar larvae.
F I G U R E 6 Formation of ventral longitudinal muscles (VLM) is affected by tup knockdown. (a-g) Show adult posterior hearts stained with phalloidin and respective orthogonal views of abdominal segment 4 (a 0 -g 0 ) and details of cardiomyocytes myofibres (a 00 -g 00 ). Control flies (a-a 0 ), as well as 76E11-Gal4-(b-c 0 ), and handC-Gal4-(f-g 0 ) driven tup knockdown animals possess prominent VLMs in close association to the heart. However, mesodermal knockdown of tup impaired the formation of VLMs (d-e 0 ). Alary muscles (am) were present in all examined animals. The yellow line in (a-g) indicates the position from orthogonal views (a 0 -g 0 ). The red-colored line indicates the adult heart tube, and the green-colored line shows the VLM. was used. The expression patterns of Gal4-driver lines used herein were previously described: hand enhancers (Sellin et al., 2006), mef2 enhancers (Gajewski, Choi, Kim, & Schulz, 2000;Gajewski, Kim, Choi, & Schulz, 1998;Nguyen & Xu, 1998), toll enhancer (Lammers et al., 2017) and 76E11 (Morin-Poulard et al., 2022) and this article.

Animal preparation and immunohistochemistry
For microscopy imaging, third-instar larvae were pinned onto Sylgard 184 silicone elastomer plates filled with PBS buffer and dissected from the ventral side. After removing viscera and respective antibody staining, specimens were embedded in Fluoromount and imaged either with a Zeiss Pascal 5 or Zeiss LSM 800 laser scanning microscope. Image processing and adjustment was performed with Affinity Photo (Serif Europe Ltd.) or ImageJ (Schindelin et al., 2012). Primary
In addition, protein levels were significantly reduced in larval aorta cardiomyocytes as shown by fluorescence intensity measurements ( Figure S3). Our controls confirmed the functionality of the two RNAi lines used in this study. For valve cell measurements, six cells per genotype were investigated. Data were analyzed using an unpaired two-tailed Student's t-test.

Quantitative RT-PCR
We used (i) the tup-RNAi lines BL51763 (Bloomington) and v103585 (VDRC) to test the general knockdown efficiency and (ii) the ubiquitously active da-Gal4 driver to induce RNAi hairpin expression. Total-RNA isolated from third-instar wandering larvae (RNeasy Mini Kit, Qiagen, Hilden, Germany) was treated with DNase 1 (Invitrogen, Life Technologies, Carlsbad, CA) according to the manufacturer's instructions and used as a template for cDNA synthesis (Luna Script Reverse Transcriptase, New England Biolabs, Ipswich, MA). We performed qRT-PCR according to standard protocols using "ORA qPCR Green ROX L Mix" (highQu, Kraichtal, Germany) and an iCycler iQ Real-Time PCR System (Bio-Rad, Munich, Germany). Primer pairs were designed with "QuantPrime" applying the pre-settings to consider only regions containing at least one intron and to accept splice variant hits. Data were evaluated as described elsewhere (Simon, 2003). The rp49 gene was used as the reference. Primer pairs used were 5 0 -cacaaatggcgcaagcccaag-3 0 (rp49 forward), 5 0 -cattttttaactaaaagtccg-3 0 (rp49 reverse), 5 0 -AACAGCTGCATACGCTCAGAAC -3 0 (tup forward) and 5 0 -AGCGCTTGTTCTGGAACCATACTC-3 0 (tup reverse). At least three biological replicates, consisting of three technical replicates each, were performed.