Single-cell expression profile of Drosophila ovarian follicle stem cells illuminates spatial differentiation in the germarium

Dong, Zhi; Pang, Lan; Liu, Zhiguo; Sheng, Yifeng; Li, Xiaoping; Thibault, Xavier; Reilein, Amy; Kalderon, Daniel; Huang, Jianhua

doi:10.1186/s12915-023-01636-9

Research article
Open access
Published: 20 June 2023

Single-cell expression profile of Drosophila ovarian follicle stem cells illuminates spatial differentiation in the germarium

Zhi Dong¹^na1,
Lan Pang¹^na1,
Zhiguo Liu¹,
Yifeng Sheng¹,
Xiaoping Li²,
Xavier Thibault³,
Amy Reilein³,
Daniel Kalderon ORCID: orcid.org/0000-0002-2149-0673³ &
…
Jianhua Huang¹

BMC Biology volume 21, Article number: 143 (2023) Cite this article

2077 Accesses
2 Citations
3 Altmetric
Metrics details

Abstract

Background

How stem cell populations are organized and regulated within adult tissues is important for understanding cancer origins and for developing cell replacement strategies. Paradigms such as mammalian gut stem cells and Drosophila ovarian follicle stem cells (FSC) are characterized by population asymmetry, in which stem cell division and differentiation are separately regulated processes. These stem cells behave stochastically regarding their contributions to derivative cells and also exhibit dynamic spatial heterogeneity. Drosophila FSCs provide an excellent model for understanding how a community of active stem cells maintained by population asymmetry is regulated. Here, we use single-cell RNA sequencing to profile the gene expression patterns of FSCs and their immediate derivatives to investigate heterogeneity within the stem cell population and changes associated with differentiation.

Results

We describe single-cell RNA sequencing studies of a pre-sorted population of cells that include FSCs and the neighboring cell types, escort cells (ECs) and follicle cells (FCs), which they support. Cell-type assignment relies on anterior–posterior (AP) location within the germarium. We clarify the previously determined location of FSCs and use spatially targeted lineage studies as further confirmation. The scRNA profiles among four clusters are consistent with an AP progression from anterior ECs through posterior ECs and then FSCs, to early FCs. The relative proportion of EC and FSC clusters are in good agreement with the prevalence of those cell types in a germarium. Several genes with graded profiles from ECs to FCs are highlighted as candidate effectors of the inverse gradients of the two principal signaling pathways, Wnt and JAK-STAT, that guide FSC differentiation and division.

Conclusions

Our data establishes an important resource of scRNA-seq profiles for FSCs and their immediate derivatives that is based on precise spatial location and functionally established stem cell identity, and facilitates future genetic investigation of regulatory interactions guiding FSC behavior.

Background

Adult stem cells provide a lifelong source of specific differentiated cells, necessitated by physiological turnover or changing environmental conditions [1,2,3]. The study of adult stem cells is revealing that this task can be accomplished in a variety of ways. Importantly, it is often a shared task. In such paradigms, often termed population asymmetry, individual stem cells make stochastic decisions about cell division and differentiation [2,3,4,5,6,7,8]. Division and differentiation are independent processes in the three best-studied paradigms [9,10,11] and are balanced at the community level. The stochastic variation in behavior of individual stem cells may be compounded further by systematic heterogeneity based on precise stem cell location [2, 12]. For example, within a population of about 16 follicle stem cells (FSCs) in a Drosophila germarium, roughly half are in a posterior ring. Only posterior FSCs can directly become proliferative follicle cells (FCs) and they also divide much faster than their anterior FSC neighbors, which can directly become quiescent escort cells (ECs) [13]. Posterior and anterior FSCs nevertheless form a single community because they can exchange positions [13]. In other paradigms, it has also often been observed that stem cell derivatives can revert to stem cell status under physiological or stress conditions [14,15,16]. All of these plastic properties, including variable stem cell lifetimes, differ substantially from the original concept, still relevant to some paradigms, of each stem cell behaving the same way—dividing with irreversible asymmetric outcomes for the two daughters and exhibiting exceptional longevity [1,2,3, 7].

Many aspects of the status of a cell can be captured by a detailed rendition of its pattern of RNA expression, so understanding of adult stem cell biology surely benefits from such information, typically captured by single-cell RNA (scRNA) sequencing. However, adult stem cell paradigms that exhibit stochastic and heterogeneous behaviors, with derivative cells that are necessarily initially very similar to stem cells, present a significant challenge to categorization of stem cells and their immediate derivatives according to scRNA profiles. These profiles cannot be expected to reveal stem cells as a highly distinct group and certainly cannot suffice to define stem cells. Instead, they must be related carefully to the results of functional tests that define stem cells through their behavior [2]. Drosophila FSCs present a particularly attractive paradigm for such analyses because their behavior is complex but has been studied carefully, along with extensive investigation of key regulatory niche signals [2, 9, 13, 17, 18]. Here, we present a scRNA study of somatic ECs, FSCs, and early FCs within the Drosophila germarium and we relate those studies to functional tests of FSC behavior based on location. The resulting expression profiles reveal candidate positionally regulated effectors of FSC division and differentiation.

Results

Location of FSCs

At the anterior of each germarium are non-dividing somatic terminal filament (TF) and cap cells, which contact 2–3 germline stem cells (GSCs) (Fig. 1A). GSCs generally divide asymmetrically to produce a more posterior cystoblast, which divides four times with incomplete cytokinesis [19,20,21]. The developing germline cyst derivatives are wrapped by processes of somatic escort cells (ECs) as they progress posteriorly through region 1 [22, 23] to form a rounded 16-cell stage 2a cyst, which then elongates into a lens-shaped 2b cyst, spanning the widest part of the germarium (Fig. 1A). Proliferative follicle cell precursors (“FCs”) associate with the 2b cyst, which then rounds to become a stage 3 cyst before budding from the germarium as an egg chamber, enclosed in a monolayer epithelium of FCs. Quiescent polar cells at the anterior and posterior termini of the FC epithelium contact non-dividing stalk cells, which connect consecutive egg chambers. Polar and stalk cell precursors are specified early within the germarium [24,25,26], while other FCs continue to divide until mid-oogenesis (stage 6), later specializing according to their anteroposterior (AP) and dorsoventral (DV) locations on the egg chamber [27, 28]. FSCs lie between ECs and the earliest FCs (Fig. 1A), providing a continuous supply of new FCs and also, less frequently, replenishing ECs [13, 25].

FSCs, like any adult stem cells, are defined by the key functional criterion of behavior over time. Lineage assays revealed the presence of about 14–16 FSCs per germarium, distributed around the inner surface of the germarium, principally in two rings of cells (“layer 1” and “layer 2”) immediately anterior to the key landmark of the anterior border of strong surface Fasciclin 3 (Fas3) staining (Fig. 1A) [13, 17, 29]. Assigning cell locations relative to a consistent landmark is a challenging task because the three-dimensional structure of individual germaria can be irregular and a dynamic process is being sampled at random times throughout a 12 h cycle of FC recruitment. We therefore describe a detailed protocol (see “Methods”) to promote consistency for all investigators and provide numerous examples from the lineage studies summarized below (Fig. 1B–K).

The location of FSCs was inferred by examining a large number of marked FSC lineages and using those with only a single-candidate stem cell to determine FSC locations [2, 13]. This strategy indicated that FSCs were almost all either immediately anterior to the strong Fas3 border (layer 1; white arrows in Fig. 1C, E, F, H, I) or one cell further anterior (layer 2; cyan arrows in Fig. 1D, G, I, J), with a few FSCs in the next most anterior location (layer 3; yellow arrow in Fig. 1J). The total average number of layer 1 (eight) and layer 2 cells (six), distributed in radially equivalent locations abutting the germarial wall (Fig. 1K), was consistent with estimates of total FSC numbers derived from (i) measuring the average contribution of a single FSC lineage to all FCs and (ii) counting the number of different multicolored lineages in a single ovariole (in each case, taking steps to ensure capturing the full diversity of all FSC lineages, including the most short-lived) [2, 13]. After defining FSC locations in this way, it was then possible to count labeled FSCs in layers 1 and 2 explicitly in numerous studies exploring the effects of altered genotypes, regardless of the number of marked FSCs per germarium [17]. Samples in such studies typically show marked FSCs in more than one layer (surviving lineages can amplify FSCs and FSCs can move between layers), several marked FCs and a limited number of marked ECs (FSCs can become FCs or ECs) (Fig. 1B). Changes in the measured numbers and behaviors of marked FSCs due to altered genotypes yielded a self-consistent picture of FSC responses to different signaling environments [17]. Moreover, recent measurements of absolute rates of FSC division matched conversion rates to FCs and ECs [30], consistent with maintaining a stable population of about 16 FSCs. How FSCs differ from FCs and ECs, as well as potential regulators of FSC behaviors, can be informed and suggested by studying RNA profiles.

scRNA-seq investigation of anterior somatic germarial cells

Some scRNA studies have been conducted on whole ovarioles to capture all aspects of oogenesis [18, 31, 32]. We were interested only in examining FSCs and their neighbors in detail, so we chose a strategy that began with purifying those cells. The enhancer trap line C587-GAL4, which is expressed in adult ECs, FSCs, and the earliest FCs, was used to drive UAS-CD8-RFP cell surface protein expression, followed by manual dissection of ovaries, dissociation, and fluorescence-activated cell sorting (FACS). Single-cell RNA sequences were derived on the 10X Genomics platform with unique molecular identifiers (UMIs). Average raw sequence reads per cell were over 290, 000. After genome alignment and UMI counting, genes featured in more than three cells were included, and cells with feature counts between 100 and 4500 were retained for principal component analysis using Seurat (v4.0.2).

Following initial clustering of scRNA patterns for a total of over 1300 cells into ten clusters (Fig. 2A), we examined each group for known indicators of cell identity. We recognized small groups with features of cap cells or TF cells (selective expression of Lmx1a (LIM homeobox transcription factor 1a), engrailed (en), decapentaplegic (dpp), and wingless (wg); 43 cells) [33,34,35,36], stalk cells and their precursors (selective expression of Lamin C (LamC; also expressed in TF and cap cells) and single-minded (sim); 23 cells) [32], and germline cells (selective expression of zero population growth (zpg), chinmo (chronologically inappropriate morphogenesis), ovo, ovarian tumor (otu) and vasa; 34 cells) (Additional file 1: Fig. S1) [32]. These three cell types do not significantly express C587 > CD8-RFP, consistent with their relatively low representation (as cells escaping FACS selection). Three groups (0, 4, 9 in Fig. 2A) were not considered further on the basis of a combination of higher mitochondrial RNA content and lower total RNA counts, suggesting the possibility of damage or stress (Fig. 2B). The remaining four groups had characteristics of ECs and FSCs (Wnt4, patched (ptc), failed axon connections (fax)) [13, 37,38,39] or early FCs (Fas3, castor (cas)) [24] and were re-sorted by themselves in order to gain better resolution among somatic cells of the anterior half of the germarium. The result was a progression of six clusters, roughly consistent with progressive anterior to posterior identities (Fig. 2C, D; Additional file 2: Table 1 for Fig. 2C), as described further below.

We compared our clusters to those derived from another study that also concentrated on anterior somatic germarial cells [40] and found substantial similarities. The patterns of selectively expressed RNAs that defined groups previously labeled as “IGS1-5” (IGS, or inner germarial sheath, is an alternative name for escort cells) showed largely good correspondence to our group 0 (IGS1 and 2; Netrin A (NetA) + crocodile (croc)), group 5 (IGS3; Croc + Helical factor (Hf)), group 1 (IGS4; bangles and beads (bnb) + wunen 2 (wun2)), and group 4 (IGS5; wun2 + santa maria) (Fig. 2E–G; Additional file 3: Table 2 for Fig. 2E). Importantly, Tu et al. [40] used several of the cited “marker RNAs” for in situ hybridization to define the spatial limits of specific clusters, confirming the anterior to posterior progression of groups 0-5-1-4 in that order, as described below. Clusters 2 and 3 appear to represent more mature FCs, with ribosomal protein RNAs dominating the most upregulated genes of cluster 2 (Fig. 3E), consistent with the highly proliferative nature of FCs. This feature is presaged in the earliest FCs in group 4 by strong expression of two ribosomal protein RNAs (RpL35 and RpS6) and myc (Fig. 3D). A handful of cells within group 2 expressed JAK-STAT ligands, unpaired 1 (upd1) and unpaired 3 (upd3), characteristic of polar cells [41].

Cluster boundaries and cell identities

From RNA in situs, santa maria is mostly expressed posterior to the strong Fas3 border, with weaker expression just anterior to the Fas3 border, overlapping with the posterior margin of bnb RNA [40]. Santa maria RNA is prevalent in group 4 but almost completely absent from group 1, while bnb RNA is highest in group 1 but present in some cells of group 4, suggesting that the key Fas3 border is roughly between groups 1 and 4, with most FSCs in group 1 (Fig. 2E, G). A few of the most posterior FSCs, expressing both bnb and santa maria, may be in group 4.

bnb expression was strongest in group 1 but also significant in abundance and prevalence in group 5 (Fig. 2E, G). Since the anterior boundary of bnb RNA expression in situ appears to extend beyond FSCs into region 2a (r2a) ECs [40], the transition between ECs and FSCs is likely very close to the border between groups 5 and 1. Other genes with sharp declines from group 5 to 1 include Hf and CG34417 (Fig. 3B), with the converse pattern for CG42797 and frizzled 2 (fz2) (Fig. 3C). Several RNAs peak within group 1, including headcase (hdc), Kank, sallimus (sls), prominin-like (promL), and sprouty (sty)) (Fig. 3C).

Regarding the more anterior clusters, NetA RNA is restricted to within region 1, while the posterior border of croc RNA extends into region 2a, declining quite sharply at the anterior border of bnb [40]. NetA expression was limited to group 0, while croc RNA was present also in a minority of group 5 cells, suggesting that groups 0 and 5 largely represent region 1 and 2a ECs, respectively (Fig. 2E, G).

Considering only groups 0, 5, and 1 (478 cells), the total number of cells in groups 0 (231, 48%), 5 (62, 13%), and 1 (185, 39%) are roughly in proportion to the average number of region 1 (r1) ECs (26, 46%), r2a (14, 25%), and FSCs (16, 29%) in an adult germarium [13, 42, 43], suggesting rough equivalence if all cells were captured equally well for sequencing. Thus, group 1 cells appear to correspond largely to FSCs, though it is expected that a few FSCs may also be within group 4 and not all cells in group 1 are necessarily FSCs (there may be some r2a ECs and some FCs) (Fig. 2D).

Prospective FSC labeling for lineage analyses

The previous section explained how a specific cluster was assigned FSC status based on the spatial expression profile of various RNAs relative to a landmark (the anterior border of strong Fas3 expression), which had been used to define FSC locations in functional lineage tests. The single-candidate method, described earlier, used to identify FSC locations is effective but it has not commonly been used for other stem cells. Instead, for mouse stem cells, the traditional approach has been to find a sufficiently specific marker to label a subset of cells through a recombination event and then determine by lineage analysis if such labeled cells have stem cell properties [2, 44]. In practice, there is often no single suitable marker with absolute specificity. Instead, an empirical threshold of marker gene expression is engineered to initiate a Cre-loxP recombination event under specific conditions of sensitivity. Consequently, questions inevitably remain about whether all cells captured by that strategy are stem cells and whether only a subset of stem cells have been captured [2]. These limitations are not commonly voiced but are significant. Despite the limitations of this targeted approach, it could usefully be applied also to FSCs, provided suitably specific targeting of cells in specific locations can be accomplished. We tried to use this approach to provide further confirmation of the location of FSCs.

The most common targeting strategy in Drosophila uses specific gene regulatory elements to drive yeast GAL4. GAL4 then activates a UAS-flp recombinase gene to initiate stable GFP expression in a recombinant cell and its derivatives (“G-trace”) [45]. Temperature-sensitive GAL80, which inhibits GAL4, is generally included, so that the timing of recombination events can be limited to a chosen time window at the restrictive temperature. The activity of a specific GAL4 transgene does not necessarily reflect normal expression of the parent gene and is normally assayed by including a UAS-RFP reporter. From scRNA-seq studies, RNA in situs, enhancer trap lines and other measures of patterned gene activity, a handful of GAL4 lines have been identified as having potentially suitable spatial specificity. According to UAS-RFP patterns, specific Wnt4-GAL4, C587-GAL4, and fax-GAL4 lines have been reported to be expressed largely or entirely anterior to the strong Fas3 border [46]. We found that after at least 3 days at the restrictive temperature, RFP expression was mostly anterior to the strong Fas3 border for all three lines, with strongest expression for Wnt4-GAL4 and weakest for fax-GAL4 (Fig. 4A–F). All Wnt4-GAL4 germaria (24/24) included RFP-positive cells anterior to the Fas3 border in locations we have previously designated as layer 1–3 FSCs, compared to 19/27 (70%) for fax-GAL4. Most such cells were labeled in each Wnt4-GAL4 germarium but only an average of 2.3 (44 total) for fax-GAL4. There was also detectable RFP in cells immediately posterior to the Fas3 border in about a third (8/24) of germaria for Wnt4-GAL4 (generally one cell with weak expression). RFP label was detected posterior to the Fas3 border in very few (3/59) fax-GAL4 germaria. For C587-GAL4, RFP was detected after 7 days at 30 °C in both layer 1–3 locations anterior to the Fas3 border and at least one cell posterior to the border in all cases (10/10). In the same samples, GFP patterns roughly mirrored RFP expression. However, while RFP expression provides an analog measure of GAL4 activity, GFP expression is digital, with permanent expression at a fixed level (driven by a ubiquitin promoter) once expression is triggered by a recombination event. Moreover, the threshold for triggering a stochastic recombination event can be quite low compared to generating a clear RFP signal. Hence, for lineage studies it is imperative to ascertain the initial GFP labeling pattern and not sufficient to rely on RFP expression patterns.

Previous G-trace tests used long periods at the restrictive temperature and reported that many ovarioles later included GFP-labeled FCs for Wnt4-GAL4 and C587-GAL4 [46]. Those results are broadly consistent with FSCs being located anterior to the strong Fas3 border. However, a similar experiment using fax-GAL4 was reported to produce very few ovarioles with labeled FCs under normal conditions in well-fed flies [46]. This evidence was used to argue that FSCs lie posterior to the strong Fas3 border. We repeated those conditions for fax-GAL4 and found substantially different results. GFP-labeled FCs were seen in 34% (17/50) of ovarioles; each of these ovarioles also included labeled cells in layer 1–3 (with an average of close to 3 labeled layer 1–3 cells per germarium), while only 9/50 included labeled cells immediately posterior to the Fas3 border. These results, in contrast to the previous report [46], are consistent with layer 1–3 cells as a source of FCs and do not appear to be consistent with the suggestion of long-lived FSCs located posterior to the AP border. Under the same conditions, C587-GAL4 (10/10 = 100%) and Wnt4-GAL4 (24/24 = 100% even by 3d) produced a greater frequency and density of FC labeling, but these two GAL4 lines are expressed at higher levels and in more cells anterior to the Fas3 border than fax-GAL4.

A much better lineage test would be to label single or very few cells in each germaria over a short time period, ensure no further labeling, and then “chase” for a period sufficient to determine if labeled FCs are formed. We experimented with a variety of conditions to achieve this outcome. A significant practical limitation here (and elsewhere [42, 47]) was the capacity of the temperature-sensitive GAL80 transgenes to silence GAL4 activity at the permissive temperature. Even in the presence of two such transgenes, GFP clones were occasionally initiated for both Wnt4-GAL4 and fax-GAL4 for flies raised and aged at 18 °C. The nature of those GFP labeling patterns was variable but the frequent heavy-labeling of multiple ECs, FSCs, and FCs suggested that most recombination events occurred during development [42], when the expression patterns of Wnt4-GAL4 and fax-GAL4 are potentially significantly different from those in adults. C587-GAL4 with three copies of the ts-GAL80 transgene gave virtually no background GFP-positive clones at 18 °C. A second experimental limitation concerns timing prior to determining initial GFP-labeling patterns. For the weakest GAL4 line, fax-GAL4, it was necessary to incubate for 10.5 h at the restrictive temperature to have relevant cells labeled in at least 20% of ovarioles. Ideally, flies are then returned to 18 °C for long enough for all recombination events to manifest as detectable GFP signal before assaying a cohort of flies for initial (“time zero”) GFP expression patterns. However, waiting too long will allow some initially labeled cells to divide or to move to different locations, including crossing the strong Fas3 boundary. We developed near-optimal timing for each GAL4 line but the compromise between a sufficient induction period and assaying a true “time zero” was greatest for fax-GAL4 because it had the weakest expression. Conversely, the third ideal requirement of limiting GAL4 activity and initial GFP marking strictly to cells anterior to the strong Fas3 border was not met by Wnt4-GAL4 and C587-GAL4. An ideal single GAL4 line, with sufficiently strong and selective expression, may not be attainable because of the inevitable similarities between a layer 1 FSC and the FC that it can become within a few hours.

Using the best conditions we could establish, we conducted quantitative G-trace experiments for all three GAL4 lines. Samples were scored to ascertain initial GFP labeling patterns (day 1) and after a further 6 days at 18 °C (day 7). Since development is roughly twice as slow at 18 °C compared to 25 °C and egg chambers generally bud every 12 h at 25 °C, it is expected that the most immature FCs (“immediate FCs”) will progress out of the germarium (2 days) and occupy the 4th or 5th egg chamber after 6 days, as previously observed directly for a 3-day interval at 25 °C [9]. A FSC may simply remain as a FSC but if it or its progeny become a FC, such FCs would be in the 4th egg chamber or more anterior locations. Thus, based on our current understanding, we would expect that all marked FCs anterior to egg chamber 4 on day 7 are derivatives of initially marked FSCs (and a few FSC derivatives may also be present in the 4th egg chamber).

For Wnt4-GAL4, a small number of ovarioles (< 15%) at day 1 and day 7 included strong, widespread GFP labeling and presumably originated during development, as described above. These background clones were discounted and not included in totals because any additional labeling induced at the restrictive temperature in adults could not be scored. At day 1, 19/34 germaria had cells labeled in layers 1–3 anterior to the strong Fas3 border (total of 37 such labeled cells; average 1.9), and 14/34 had at least one GFP-positive cell immediately posterior to the Fas3 border (8 of these also had label in layers 1–3) (Fig. 4K, L). At day 7, the observed fraction with marked FCs anterior to the 4th egg chamber was 37% (11/30) (Fig. 4O). These results are compatible with the marked FCs at day 7 originating from either layer 1–3 cells or cells posterior to the Fas3 border. Essentially, Wnt4-GAL4 initially labels cells posterior to the Fas3 border too frequently to provide a decisive test, even though the pattern of UAS-RFP labeling is strongly biased towards cells anterior to the Fas3 border.

For C587-GAL4 at day 1, 14/58 germaria had cells labeled in layers 1–3 anterior to the strong Fas3 border (total of 20 such labeled cells; average 1.4), and 8/58 (14%) had a GFP cell immediately posterior to the Fas3 border (4 of these also had label in layers 1–3) (Fig. 4I, J). At 7 days, 10/41 (24%) of ovarioles included labeled FCs anterior to the 4th egg chamber and a further 5/41 (12%) included labeled cells in layers 1–3 with no labeled FCs (Fig. 4N). The proportion of 7-day ovarioles with labeled FCs (24%) exceeded the proportion with labeled cells posterior to Fas3 on day 1 (14%), suggesting that there must be a major source of FCs anterior to Fas3. By contrast, the proportion of germaria with labeled layer 1–3 cells at day 1 and labeled FCs at day 7 were similar, consistent with the former serving as FSCs.

For fax-GAL4, 15/123 ovarioles at day 1 and 11/81 ovarioles at day 7 showed heavily labeled background GFP signal and were excluded. At day 1, 26/108 (24%) of germaria had cells labeled in layers 1–3 anterior to the strong Fas3 border (total of 70 such labeled cells; average 2.7), and 7/108 (6%) had a GFP cell immediately posterior to the Fas3 border (6 of these also had label in layers 1–3) (Fig. 4G, H). At 7 days, 15/70 (21%) of ovarioles included labeled FCs anterior to the 4th egg chamber, FCs in the 4th egg chamber were the only GFP-positive cells other than ECs in 4/70 (6%) ovarioles, and a further 4/70 (6%) included labeled cells in layers 1–3 with no labeled FCs (Fig. 4M). The proportion of 7-day ovarioles with labeled FCs (24%) is much higher than the proportion with labeled cells posterior to Fas3 on day 1 (6%), showing that the major source of FCs must be anterior to Fas3. The frequency of ovarioles with labeled layer 1–3 cells at day 1 (24%) roughly matched the sum of FC derivatives (21%) and layer 1–3 cells with no FC derivative (6%) at day 7, consistent with the expected behavior of layer 1–3 cells as FSCs [2, 13]. FCs in the 4th egg chamber (6%) may have derived from FSCs or cells immediately posterior to the Fas3 border (6%). In summary, our examination of lineages targeted largely to cells anterior to the strong Fas3 expression border supports our earlier assertion that FSCs lie anterior to this border, and hence the assignment of group 1 from scRNA-seq studies as being centered on FSCs.

Patterns of gene expression

The patterns of gene expression emerging from scRNA-seq studies of ECs, FSCs, and FCs can provide an overview of the developmental progression among these cell types and can suggest specific genes to examine for functional roles in driving transitions from FSCs to FCs or ECs. Different displays of data are described below to suit different purposes. In all cases, the raw values of average expression of a gene (number of UMIs) per cell in each cluster were first scaled and normalized to a total of 10,000 for all 8065 genes included. Thus, each value represents the fraction of all measured RNAs in a cell representing the gene in question. This number also approximates the actual average number of counts of the corresponding RNA per cell of the cluster because the average number of UMIs per cell was 9213 (average of 1.14 over 8065 genes). The average UMI count per gene per cell was progressively higher for cells from anterior ECs (group 0; 0.8), through posterior ECs (group 5; 1.1), FSCs (group 1; 1.6) to early FCs (group 4; 2.5) (Fig. 5A). This may indicate increasing overall RNA levels but might also reflect technical differences in complete cell and RNA capture.

The normalized expression values were used first to sort genes in a manner that highlights enrichment in single clusters. Specifically, Additional file 4: Table 3 lists genes characteristic of each cluster, ranked by enrichment (“E”) in the appropriate group “X” according to the value of E = (log2 (1 + expression in X)/(1 + average expression in all other groups). Only enrichment values greater than 0.4 are listed (all values for all genes can be found in Additional file 5: Table 4), together with the percentage of cells expressing the gene within the cluster and outside the cluster, and the probability of observing this enrichment by chance. A heat-map of the top 20 genes for each cluster is presented in Fig. 5C. Heat-maps for the top 60 genes for each of the two major principal components (PC1 and PC2) used for clustering are presented in Fig. 6. Looking at all three heat-maps, the most prominent, or sharpest, transition in expression is between groups 1 (FSCs) and 4 (early FCs). This pattern is especially obvious for PC1 (Fig. 6A; Additional file 6: Table 5) but also apparent principally for genes characteristic of group 5 and group 2 (Fig. 5C), suggesting that many of the relevant genes are characteristic of EC and FC function, respectively. The genes that are activated in group 2 (FCs) and PC1 are dominated by ribosomal protein genes, and also include genes encoding products involved in ribosome assembly (stubarista; sta) and DNA replication (Proliferating Cell Nuclear Antigen; PCNA). This is consistent with the designation of groups 4 and 2 as highly proliferative FCs, groups 0 and 5 as non-proliferating ECs, and group 1 as FSCs with variable and relatively low proliferation rates.

The second-most striking transition, highlighted by PC2 (Fig. 6B; Additional file 7: Table 6) and evident for genes characteristic of groups 0, 5, 1, and 4, is between groups 5 (posterior ECs) and 1 (FSCs) (Fig. 5C). Genes that are highly expressed in ECs (group 5 and group 0) and reduced in group 1 (FSCs) are generally further reduced in FCs of groups 4 and 2. However, expression of some of the genes that are largely repressed in ECs and activated in FSCs (group 1) decline in group 4 and even more prominently in group 2 FCs, suggesting that some aspects of the EC to FSC transition are dissimilar to those of FSC to FC transitions. These patterns, emphasized by PC1 and PC2, align with previous in situ patterns of gene expression assayed by reporters and antibody studies, showing several genes expressed in ECs, declining in FSCs and terminating close to the first FCs (including the Wnt pathway activity reporter, Fz3-RFP, as well as the PZ1444-lacZ and C587-GAL4 enhancer traps), but none with a sharp transition between ECs and FSCs [13, 32, 40]. These same characteristics are also evident when examining the same data as displayed in the heat-maps as “dot plots”, which additionally display visually the percentage of cells with detected transcription of each gene (Fig. 3). Several genes characteristic of group 4 or 2 (FCs) have comparatively low expression in group 1 (FSCs) (Fig. 3D, E), whereas most genes characteristic of groups 5 and 0 (ECs) do not decline greatly in FSCs (group 1). A couple of exceptions, which may be interesting to pursue functionally, are jitterbug (jbug), encoding a Filamin-type protein [48] and Helical factor (Hf), encoding a secreted cytokine. This pattern of Hf expression was noted and shown by RNA in situs in a prior scRNA-seq study [40]. Whether examined by heat-maps (Fig. 5C and Fig. 6), t-SNE plots (Fig. 5B), or dot plots (Fig. 3), strictly FSC-specific (group 1) expression is not evident. Prominently expressed genes are also expressed significantly either in ECs or FCs.

In an adult ovary, the key regulated behaviors of FSCs are division and differentiation. Differentiation at the anterior face of the FSC domain is to ECs, while posterior FSCs directly give rise to FCs. Both differentiation events are influenced by the strength of Wnt signaling and JAK-STAT signaling [13, 17, 38], but the mediators of these two differentiation events are unknown and may be quite different. Wnt pathway activity declines from anterior to posterior over the FSC domain, while the JAK-STAT signaling gradient has the opposite polarity [13, 17, 38, 49]. Some mediators of these signals might therefore be expected to show graded expression from ECs through FSCs to FCs, with either polarity. However, this straightforward correspondence is likely often modified because some mediators may be dedicated to regulating only one transition of FSCs and patterns of gene expression of potential mediators are likely influenced by additional signals. It is therefore of interest to display scRNA-seq data in a manner that highlights genes with significantly graded expression between ECs and FSCs, between FSCs and FCs, or both. Some of these genes may be responsive to Wnt or JAK-STAT signals and may directly regulate division rate or differentiation to ECs or FCs. The spreadsheets in Additional files 8 and 9 report fractional increases or decreases, always relative to the larger number, for transitions from ECs to FSCs and from FSCs to FCs. Only genes with an average expression value per cell greater than 1.5 in at least one group were included to focus on the most robustly supported changes. Genes are ranked according to the magnitude of these differences (increases from anterior to posterior in Additional file 8: Table 7, decreases in Additional file 9: Table 8, opposing changes between the two transitions in both spreadsheets), featuring only the top two quartiles (color-coded) for each transition (values for all genes can be found in Additional file 5: Table 4). Gene order within each set was determined by the magnitude of fractional changes after combining groups 0 and 5 for ECs, and combining groups 4 and 2 for FCs. The same calculations are reported and color-coded for transitions from groups 5 to 1 and from groups 1 to 4. The color-coded patterns for EC to FSC and FSC to FC transitions are reproduced in Fig. 7A. Since there are two top quartiles for increases (“UP”) and decreases (“DOWN”), and four lower quartiles for increases together with decreases (“NEUTRAL”), random association of changes from EC to FSCs and FSCs to FCs would result in 25% of UP for one transition being associated with UP or DOWN for the second transition (50% associated with NEUTRAL). There are some major departures from this pattern. Naming changes in the EC to FSC transition first, the frequencies of UP-UP (50/379; 13%) and DOWN-UP (42/416; 10%) are very low, while the frequency of UP-DOWN is very high (184/379; 49%) compared to the default expectation of 25% frequency. The much higher frequency of UP-DOWN (184/379; 49%) compared to UP-UP (13%) indicates that the changes between EC and FSC states differ substantially from those during a FSC to FC transition.

Some of the genes that are regulated principally by Wnt or JAK-STAT signaling and involved in promoting the FSC to FC transition might be expected to fall in the UP-UP category. Within this relatively small category (50 genes) are ten genes involved in ribosome biogenesis (rRNA processing: Nop60B (Nucleolar protein at 60B), Fib (Fibrillarin), Nop56; 60S biogenesis: Non1 (Novel nucleolar protein); ribosomal proteins: RpL35, RpL5, RpS19a) or protein translation (translation initiation factors eIF1b, eIF1d, eIF3a), another two linked to DNA replication and proliferation (PCNA, dUTPase (deoxyuridine triphosphatase)), a Myc collaborator (modulo; mod), genes related to nuclear transport (Pendulin; Pen, Ran) and to fatty acid oxidation (Enoyl-CoA hydratase, short chain 1; Echs1). Of the three transcription factors in this group (cut; ct, castor; cas, hairy; h), Cut contributes to FC proliferation [50] and Castor is a key marker of early FCs, which later collaborates with eyes absent (Eya) to specify different types of FC [24]. Finally, Fas3, like Cas, is a key marker that increases from FSCs to FCs, while insulin-like peptide (Ilp6) is potentially important to stimulate insulin/PI3-kinase pathway activity, which is known to affect FSC division rates [51].

Potential regulators of FSC behavior

To identify further potential regulators of FSC transitions, we prioritized the nature of the gene rather than the absolute hierarchy with regard to magnitude of changes, or comparison of changes from EC to FSCs versus FSCs to FCs. The process of differentiation of FSCs to ECs or FCs is not well understood but it certainly involves cell movement along the AP axis. Adhesion proteins are therefore likely important mediators, potentially promoting posterior FSCs joining a nascent FC epithelium, anterior FSCs moving towards ECs, or FSCs moving along a gradient of extracellular matrix (ECM) adhesion ligands. We therefore looked for adhesion molecules (and some partners and regulators) with a clear gradient of expression from group 5 through group 1 to group 4 (Fig. 7B, C). Among cell–cell adhesion molecules with increasing expression from ECs (group 5) to FCs (group 4) were Fas3, DE-cadherin (shg), and immunoglobulin superfamily genes, Plum and babos. The opposite pattern was identified for Fasciclin I (Fas1), echinoid (ed), Neuroglian (Nrg), Tenascin accessory (Ten-a), connectin (con), and capricious (caps). Among these genes, only the role of DE-cadherin in FSCs has so far been investigated [52]. An increase in expression from posterior ECs (group 5) to FSCs (group 1), followed by a decline in FCs (group 4) was seen for three ECM components (Collagen type IVa1 (Col4a1) and the Laminin subunits, wing blister (wb) and LanB1), an integrin β subunit (myospheroid, mys), and the major non-integrin ECM receptor (Dystroglycan (Dg).

A subset of genes encoding proteins involved in organizing adhesion complexes or connecting them to the actin cytoskeleton is illustrated as “adhesion-related” (Fig. 7C), and grouped according to prior evidence of involvement in septate junctions (coracle (cora) and Lachesin (Lac)), ECM interactions (Dystrophin (Dys) and parvin), microdomain-clustering Tetraspanins (Tsp74F, Tsp42Ee, Tsp42Ea, Tsp66E), Cadherin collaborators (a-catenin (a-cat), Adherens junction protein p120 (p120ctn), raskol GTPase and Innexins, Inx 2 and Inx3), other adherens junction proteins (bazooka (baz), crumbs (crb), canoe (cno) and Zyxin (Zyx)), and Filamin-family actin cross-linkers cheerio (cher) and jitterbug (jbug). Genes within most of these categories show opposing patterns of expression but Cadherin collaborators all decreased from ECs through FSCs to FCs, in contrast to DE-cadherin (shg) itself.

Several of the adhesion molecules described above are used for axon guidance (including Fas3 and Fas1, Nrg, Caps, and Con). Significant changes in gene expression were also found for guidance molecules not directly involved in cell adhesion (Fig. 7D), including a Netrin ligand (NetA), two Semaphorin ligands (Sema1a, Sema2a), two receptors for a repulsive Sema 1a signal (off-track (otk) and PlexinA (PlexA)), and a receptor for Slit ligand (roundabout 2 (robo2)).

Posterior FSCs cycle faster than anterior FSCs, while ECs do not divide at all [13]. Graded JAK-STAT activity, declining from the posterior, is one contributor to this patterned proliferation, while Integrin, Hedgehog (Hh), and Phospho-inositide 3’ kinase (PI3K) pathways also provide major stimuli for FSC division [17, 51, 53,54,55]. Most ECs are found in G1, while most posterior FSCs are in G2, indicating limitation of cycling by G2/M restriction in all FSCs with an increasingly severe G1/S restriction further anterior [30]. We found that several RNAs associated with DNA replication increased from anterior (groups 0 and 5) to posterior (group 4), including Proliferating Cell Nuclear Antigen (PCNA), deoxyuridine triphosphatase (dUTPase), Ribonucleotide reductase small (RnrS), and large (RnrL) subunit genes. A similar, graded increase in drivers of mitosis was seen, including fizzy (fzy), polo, cyclin B (cycB), and cyclin B3 (cycB3), though string cdc25 phosphatase (stg) had strongly graded expression of opposite polarity (Fig. 7E).

Signaling pathway activities can be indicated by the magnitude of expression of ligands, although the range of ligand movement can be variable and regulated. Less frequently, pathway activity is dependent on the expression level of receptors, which are also sometimes upregulated or downregulated in response to signal. Additionally, some genes are known to respond transcriptionally to pathway activity, often as part of a negative feedback circuit, but such expression is sometimes tissue-specific, rather than universal. These gene categories that were found to have significantly patterned expression are grouped according to the signaling pathway in Fig. 7F. Insulin-like peptide 6 (Ilp6) expression increased prominently towards the posterior, with insulin receptor (InsR) expression fairly uniform over the critical group 5-1-4 domain. The activin-like agonist dawdle (daw) was notably high in group 0 ECs, while expression of the activin antagonist Follistatin (Fs) [56] increased either side of the FSC region. BMP ligands (decapentaplegic (dpp), glass bottom boat (gbb)), two genes commonly induced by pathway activity (Dads against dpp (Dad) and brinker (brk)) [57], and the receptor thickveins (tkv), all decreased towards the posterior. BMP signals supplied to GSCs are strictly limited to the anterior of the germarium [58]. Hedgehog (hh) ligand, the pathway’s transcriptional effector (cubitus interruptus (ci)) and two genes generally induced by the Hh pathway, roadkill (rdx) and the Hh receptor patched (ptc) [59], were also all reduced towards the posterior. This is consistent with prior detection of hh expression in anterior germarial cells, supplementing very strong expression in cap cells and terminal filament cells, and a slow decline of pathway reporter activity in the posterior half of the germarium [60, 61]. A number of Wnt ligands (Wnt6, Wnt4), one receptor (frizzled, fz), and two genes often induced by Wnt-β -catenin pathway activity (naked (nkd) and Notum) [62, 63] also declined from anterior to posterior over the whole domain extending from ECs to FCs, while another ligand (Wnt2), receptor (frizzled 2 (fz2)), and facilitator of Wnt secretion (wntless (wls)) only declined prominently from FSCs (group 1) to FCs (group 4). Wnt6 has previously been found to be expressed only in anterior ECs, while Wnt4 is strongly expressed in all ECs, tapering off at the FSC border [39, 43, 61]. The RNA expression pattern profiles described are consistent with the observed anterior to posterior decline of a Wnt pathway reporter over the FSC domain [13, 49]. Epidermal growth factor receptor (EGFR), one of its ligands vein (vn), rhomboid-4 (rho-4), and rhomboid (rho), which process EGFR ligands including Spitz, all decline from group 5 through group 2 (Fig. 7F). However, the same pattern is seen for gone early (goe) which can attenuate EGFR signaling [64], and a gene often induced by EGFR pathway activity, kekkon1 (kek1) [65], does not decline from ECs to FSCs. Two fibroblast growth factor (FGF) ligands, pyramus (pyr) and thisbe (ths), decline from group 5 to group 1, while the FGF antagonist sprouty (sty) undergoes the converse change. The majority of genes involved in stabilization and spread of multiple ligands through protease activity (Matrix metalloprotease 1 (Mmp1)), as a glypican (dally-like protein (dlp), as heparin sulfate proteoglycan (HSPG) proteins (Syndecan (Sdc) and Perlecan, terribly reduced optic lobes (trol), HSPG associated proteins (shifted (shf) or contributing to their biosynthesis (sulfateless (sfl)) decreased consistently from anterior (ECs) to posterior (FCs). Among these, Mmp1, dlp, and sfl have specifically been associated with Wnt signaling [49, 66], dlp and shf with Hh signaling [67, 68]. Altogether, the expression patterns of several genes derived from these scRNA-seq studies are in accord with existing evidence of an anterior to posterior decline of some signaling pathways (BMP, Hh, Wnt) and provide some new indications of regionalized activity for others (activin, EGF, FGF), which may motivate further functional investigation.

Independent assessment of graded expression patterns suggested by scRNA-seq

We used available reagents for immunohistochemical examination of selected gene activities highlighted by our scRNA-seq studies as potential spatially regulated mediators of FSC behavior.

Semaphorin 1a (Sema1a) is a transmembrane protein that can bind to PlexA receptors on adjacent cells to counter cell adhesion in the context of axon guidance [69]. However, Semaphorin family proteins have been found to affect the migration of several other cell types, including Drosophila FCs [70]. A protein trap insertion showed expression initiating close to the EC/FSC interface and continuing into early FCs and then declining, decorating the long membrane extensions of FSCs and early FCs (Fig. 8A). This pattern strongly resembles that inferred from scRNA-seq (Fig. 7), which also predicts a largely overlapping pattern for PlexA. It will be interesting to compare Sema1a and PlexA protein distributions in the same samples to indicate whether their interactions might contribute to segregation of specific pairs of cells.

Dystroglyan, along with integrin, is a major receptor for ECM proteins and is typically found on the basal surface of epithelial cells [71, 72]. We found Dystroglycan protein to line the surface of the germarium, adjacent to the basement membrane in the most posteriorly located ECs, all FSCs, and early FCs (Fig. 8B). This pattern was very similar to the pattern of cells expressing Sema1a-GFP and the pattern inferred from scRNA-seq (Fig. 7). The pattern of Dystroglycan expression suggests it may contribute to restricting FSC to EC conversions through ECM adhesion.

Plum is an immunoglobulin superfamily protein, first functionally identified from involvement in axon pruning [73]. Plum can promote homophilic cell adhesion but may instead act by modulating TGF-β signaling in mushroom body neurons [73]. A Plum-GFP protein fusion encoded by a genomic transgene was expressed only at low levels in the anterior of the germarium but increased on the surface of FSCs and early FCs (Fig. 8C). The increase in Plum-GFP was further posterior than for Sema1a-GFP and Dystroglycan, consistent with the comparative scRNA-seq profiles (Fig. 7), but did not reflect a further increase from FSCs to FCs as suggested by scRNA-seq results. Hence, differential Plum expression is perhaps more likely to affect FSC to EC rather than FSC to FC transitions.

Recent examination of FSCs expressing fly-FUCCI reporters of cell cycle phase indicate that both G1/S and G2/M transitions can be rate limiting, with a prominent gradient of greater G1/S flux from anterior ECs through posterior FSCs [30]. Our scRNA-seq profiles suggest that a similar profile of CycB levels, promoting the G2/M transition, may also be relevant to the faster cycling of posterior FSCs than anterior FSCs. A CycB-GFP protein trap showed no significant expression in ECs and even anterior FSCs but sharply elevated expression more posterior, seen occasionally in posterior (Layer 1) FSCs and in most early FCs (Fig. 8D). Protein degradation during G1 and S phases may account for some of the observed pattern [74], but the majority of FSCs are in G2, so the absence of CycB-GFP in anterior FSCs is consistent with a deficit in the corresponding RNA. CycB, as suggested by scRNA-seq data, is therefore a prime candidate as a spatially mediator regulator that may limit FSC cycling.

The expression of santa maria has been noted in our scRNA-seq studies and others [32, 40, 46] to be highly selective within the germarium and therefore an important potential indicator of cell type. In our work, santa maria is characteristic of group 4, the earliest FCs, with significantly higher expression than in their immediate neighbors, FSCs (and also, more mature FCs). We examined the expression of santa maria by using a transgene with GAL4 linked to a 2 kb santa maria gene promoter to drive UAS-RFP [75]. The conditions were analogous to those described earlier for other GAL4 lines (29 °C for 7 days in the presence of temperature-sensitive GAL80). We found that germaria generally had few or no RFP-labeled cells (more than 80%), consistent with relatively low-level expression of this gene overall and stochastic variations in cells of equivalent positions and natures. RFP was seen most often in “immediate FCs”- those FCs that first have elevated Fas3 protein staining and were most recently produced (14 cells among 11germaria) (Fig. 8E–G). There were also a few additional, more mature neighboring FCs marked (11 cells) (Fig. 8G) and a small number of marked posterior FSCs (7 cells in total) (Fig. 8E). This evidence is consistent with the scRNA-seq data and group assignments. It also confirms that some FSCs may be captured among primarily early FCs in group 4. While such overlaps can reflect technical limitations of scRNA-seq, the variable santa maria > RFP expression patterns expose a fundamental issue; there is certain to be considerable overlap in gene expression between two cell types that are physically adjacent and regularly interconverted within a few hours under normal physiological conditions.

Discussion

Single-cell RNA sequencing has great potential for understanding key regulators of a cell state. A first step in realizing this potential is to relate single-cell profiles to cell identities. For FSCs, the key intermediate is spatial location. Here, we used cells pre-sorted to represent mainly anterior germarial cells (ECs, FSCs, and early FCs) and standard bioinformatic analyses to derive clusters guided by similarities. The clusters were notably similar to those of another study [40], facilitating assignment of group locations according to prior RNA in situ results for key indicators of those groups. The cluster profiles themselves showed a clear progression consistent with assignments along the AP axis and the placement of a key spatial landmark, the anterior border of strong surface Fas3 protein expression, roughly at the border between two groups (“1” and “4”). The anterior limit of group “1” was placed about three cells further anterior by comparison to RNA in situs, forming a cluster size of roughly the expected size to represent FSCs relative to the more anterior groups (“0” and “5”) representing ECs. Importantly, this study is the first to assign the FSC group by taking into account the most comprehensive functional investigation of FSC numbers and locations [13]. Some other studies sampled the whole ovariole and therefore lacked sufficient FSCs in their samples [31, 32, 46]; others did not acknowledge recent developments in understanding FSC biology or simply avoided an FSC designation, focusing instead on other cell types [40, 46, 47]. Here, we included images of a variety of GFP-marked FSC lineages to illustrate FSC locations in 2 or 3 AP layers anterior to the Fas3 border, as determined previously by examining lineages with only a single-candidate FSC [13]. We also included a protocol for identifying the Fas3 border because other investigators have presented inconsistent and potentially confusing approaches and summaries [46].

We also undertook a complementary approach to verify FSC locations through prospective “G-trace” labeling [45]. We obtained results consistent with our prior designation of FSCs occupying 2 or 3 layers anterior to the Fas3 border and contradicting an assertion that FSCs are posterior to the AP border [46], based principally on using the very same fax-GAL4 reagent to initiate clones. Crucially, our tests involved carefully controlled conditions that allowed production and measurement of a suitable number of GFP-labeled cells during a short labeling period, followed by a chase period to ascertain the fate of such cells. Results with fax-GAL4 and C587-GAL4 to initiate labeling were consistent with FC-producing cells lying anterior to the Fas3 border, although that conclusion was less numerically robust for C587-GAL4 because the associated initial GFP-labeling pattern was less specific than for fax-GAL4. The principal limitation of G-trace in this context was the failure of each GAL4 line to show absolute spatial specificity. A second limitation was that the shortest practical time between initiation of targeting recombination and assessing the cellular specificity of labeling is sufficiently long that some stem cells will differentiate and change location in the interim. These limitations to targeted labeling likely apply to most stem cell paradigms, especially those governed by population asymmetry. The best method for determining FSC locations therefore remains the single-candidate method, which allows single-cell resolution of FSC positions [2, 13]. The deduced locations and numbers of FSCs fit extremely well with many measurements of FSC division rate and FC production [2, 17, 30], as well as the targeted G-trace studies presented here.

The assignment of a group of scRNA expression profiles to FSCs is not precise for a number of reasons, including imperfect and limited depth of RNA profiles, latitude in spatial correspondence according to RNA in situs, and the systematic limitation that FSCs become ECs or FCs within a few hours and therefore inevitably have strong similarities to those neighboring derivatives. Nevertheless, the results of this study have the unique virtue that the approximately defined FSC group of cells is in accord with current understanding of FSC numbers and locations determined by thorough functional analyses [2, 9, 13, 17, 76,77,78]. FSC behavior is guided by a number of external signals but with prominent roles for two graded signaling pathways, Wnt, and JAK-STAT [13, 17]. RNAs that also have graded AP expression from ECs through FSCs to early FCs are therefore of special interest as candidate mediators of spatially appropriate behavior guided by these graded signals. We have highlighted such RNAs encoding potential effectors of cell movements and differentiation or cell division as candidates for further investigation of functional roles and potential transcriptional regulation by Wnt or JAK-STAT pathways.

Conclusions

We have sorted cells from the Drosophila germarium, performed single-cell RNA sequencing and grouped clusters of profiles according to similarity. Comparison of selectively expressed RNAs to previously determined RNA in situ patterns showed a clear AP progression among four identified clusters, including the identification of one cluster as overlapping FSC territory. We confirm previous allocation of FSC locations along the AP axis using lineage analyses. Specific candidate genes with graded expression around the FSC domain are highlighted as potential effectors of FSC division and differentiation instructed by graded extracellular signals. Although other recent scRNA-seq studies have examined cells in the Drosophila ovary [31, 32, 40, 46, 47], this study is unique by virtue of focusing on the FSC region and utilizing the most current functional understanding of FSC numbers, locations, and behavior [2, 13].

Methods

Single-cell suspension preparation and fluorescence-activated cell sorting (FACS)

Female flies of genotype C587-GAL4 / yw; UAS-CD8-RFP/ + were selected on the day of eclosion and maintained on rich food at 25 °C together with males for 4–7 days. More than 200 pairs of ovaries were dissected in ice-cold DPBS (Dulbecco’s phosphate-buffered saline) solution, and then gently washed twice by DPBS. The ovaries were digested in 700 µl solution containing 0.5% trypsin (Sigma, #T4799) and 0.25% collagenase (Invitrogen, #17,018–029) at 25 °C for about 20 min with gentle shaking. The dissociated cells were filtered through a 40-micron cell strainer (Falcon, #352,340), and then centrifuged at 400 g at 4 °C for 5 min. Supernatant was removed and the cell pellet was resuspended in DPBS with 0.2% BSA. Cell suspensions were sorted using a FACS Aria III sorter MoFlo XDP (Beckman-Coulter) based on RFP signal. Cells from w¹¹¹⁸ fly ovaries were used to set the negative fluorescence gate for the RFP panel. RFP-positive cells were sorted into DPBS containing 0.2% BSA, pelleted (400 g for 10 min), and resuspended in DPBS. The LUNA-FL double fluorescent cell counter (Logos Biosystems) was used to count cell number and the live/dead cell ratio. The suspension of single cells, with > 85% live cells, had a final concentration of 500 cells/µl.

Single-cell RNA sequencing (scRNA-seq)

The sorted cells were loaded on a Chromium Single cell Controller (10 X Genomics) using the Chromium Single-cell 3’ Library & Gel Bead kit v3 according to the manufacturer’s directions. Briefly, the Gel Beads containing the poly-T primer sequence were linked with a cell barcode and UMI (unique molecular identifier) and the cells are wrapped by “oil droplets” to form a GEM (gel bead in emulsion). Cells were lysed in these droplets and reverse-transcribed to form full-length cDNA sequence. After the oil droplets were broken and purified, the cDNA library was PCR amplified and ligated with sequencing primers. cDNA library quantification assays and quality check analysis were performed using an Agilent 2100 Bioanalyzer and Thermo Fisher Qubit Fluorometer. The library samples were then diluted to a 10 nM concentration and sequenced on an Illumina NovaSeq 6000 platform (Illumina, San Diego) for PE150 sequencing by Annoroad Gene Tech. Co., Ltd (Beijing, China). A total of 377,489,707 reads were obtained for the sample, with 290,154 mean reads per cell.

Raw data preprocessing

The raw sequencing data was processed with alignment, barcode assignment, and UMI counting by Cell Ranger (version 6.1.1) using the previously published pipeline [79]. The reference index for Cell Ranger was built using the Drosophila melanogaster genome (version BDGP6.32) available on the Ensembl genome database.

Single-cell RNA-seq data analysis

Cell Ranger’s raw gene count matrix was further analyzed using the Seurat (v4.0.2) R package with standard protocols [80]. Only genes expressed in more than 3 cells were kept, and only cells that have unique feature counts between 100 and 4500, and fewer than 20% mitochondrial counts (no cells were actually excluded due to high mitochondrial RNA counts) were selected for further analysis. Gene expression results were log-normalized, and then regressed on the percentage mitochondrial gene content. The top 16 principal components of principal component analysis (PCA) were used for clustering according to the Elbow Plot (resolution = 0.9). The ECs, FSC, and FCs were identified by markers of each cluster. t-SNE (t-distributed stochastic neighbor embedding) plots were used to visualize the clustering results. Mean expression values for each cluster were calculated by “AverageExpression” in Seurat. The R package “pheatmap” was used to generate heat-maps and the gene expression levels were scaled by row.

MARCM and multicolor FSC lineages

In a typical MARCM (mosaic analysis with repressible cell marker) experiment, 1–3-day-old adult Drosophila melanogaster females with the appropriate genotype (yw hs-Flp, UAS-nGFP, tub-GAL4 /yw; act-GAL80 FRT40A / FRT40A; act > CD2 > GAL4/ + or yw hs-Flp, UAS-nGFP, tub-GAL4 /yw; FRT42D act-GAL80 tub-GAL80 / FRT42D; act > CD2 > GAL4/ +) were given a single 30-min heat shock at 37 °C. Afterwards, flies were incubated at 25 °C and maintained by frequent passage on normal rich food supplemented by fresh wet yeast before dissection 6 or 12 days later. For multicolor lineages, the procedure was similar but with multiple heat shocks [13] and the final genotype of yw hs-flp / yw; tub-lacZ FRT40A FRT42B / ubi-GFP FRT40A FRT42B His2Av-mRFP. Adult female ovaries were dissected in PBS, leaving the tip of the abdomen attached, and fixed in 4% paraformaldehyde (Electron Microscopy Sciences, #15,710) for 10 min at room temperature, covered to prevent bleaching of fluorophores. The ovaries were then washed three times with 1X PBST (PBS with 0.1% Triton X-100 and 0.05% Tween 20) and blocked at room temperature in blocking solution (PBST with 10% normal goat serum) for 30 min. Primary antibodies mouse anti-Fasciclin 3 (Developmental Studies Hybridoma Bank, #7G10, 1:250) and rabbit anti-GFP (Invitrogen, #A6455, 1:1000) were added and samples were nutated for an hour at room temperature. The ovaries were then rinsed with 1X PBST three times for 10 min each before being incubated with secondary antibodies Alexa Fluor-647 goat anti-mouse (Invitrogen, #A21236, 1:1000) and Alexa Fluor-488 goat anti-Rabbit (Invitrogen, #A11034, 1:1000) at room temperature for 1 h. The ovaries were then rinsed twice with PBST and once in PBS. Each ovary pair was then broken apart in PBS on a slide and mounted using DAPI fluoromount-G mounting medium (Southern BioTech, OB010020). Ovarioles were imaged with a Zeiss LSM700 or LSM800 confocal microscope, operated in part by the Zeiss ZEN software. Examples from single-color MARCM and multicolor lineage analyses are shown in Fig. 1.

Protocol for identifying cell locations relative to landmark of anterior border of strong Fas3 staining

A variety of indicators help to ensure reproducible determination of the anterior border of strong Fas3 staining. These include germarium width, identifying the youngest stage 2b cyst and noting cell process locations. Starting with a mid-z-section, the youngest stage 2b cyst is identified by the criteria of spanning the germarium and not being at all rounded. Most germaria have only a single-candidate 2b cyst with a clearly more rounded stage 3 cyst more posterior. In such germaria, the Fas3 border runs along the posterior surface of the 2b cyst. In other germaria (up to about a quarter), a new 2b cyst has formed as a more posterior 2b cyst just starts to round. Here, the strong Fas3 border lies between the two 2b cysts. The strong Fas3 border can then be followed as a continuous surface through neighboring z-sections. Layer 1 cells immediately anterior to the Fas3 border have strong Fas3 staining on their posterior surface but weak or incomplete outlining of the anterior surface by Fas3. Labeling in MARCM clones usually allows visualization of FSC processes (even faintly when using a nuclear-targeted GFP marker). Layer 1 cell processes are present along the Fas3 border, whereas layer 2 processes are anterior to the stage 2b cyst. The widest part of the germarium is generally very close to the Fas3 border, with layer 1 FSCs at the widest location in about three quarters of samples and immediate FCs at that location in the rest.

G-trace experiments

Flies of genotype UAS-RFP, UAS-flp, ubi > stop > GFP / CyO; tub-GAL80(ts) / TM2 were crossed to (a) tub-GAL80(ts) / CyO; fax-GAL4 / TM2 or (b) Wnt4-GAL4 / CyO; tub-GAL80(ts) / TM2 or (c) C587-GAL4; tub-GAL80(ts) / CyO; tub-GAL80(ts) / TM2 flies at 18 °C to collect experimental progeny with no balancer chromosomes and two (for fax-GAL4 and Wnt4-GAL4) or three (for C587-GAL4) copies of tub-GAL80(ts). Young (1–4 days old) experimental flies were then shifted to 29 or 30 °C for various periods and, in some cases, returned to 18 °C, as described in the “ Results” section for specific tests. For the final lineage tests, the conditions were 3 h at 30 °C followed by 13 h at 18 °C (Wnt4-GAL4), 6 h at 30 °C followed by 13 h at 18 °C (C587GAL4), or 10.5 h at 30 °C followed by 13 h at 18 °C (fax-GAL4) before dissecting for the first time-point, followed by a further 6 days at 18 °C before dissecting another cohort. An equivalent cohort of flies was kept at 18 °C throughout. In all cases, flies were transferred to fresh food every 2–3 days. Dissection, staining, mounting, and analyses were as described for MARCM experiments (RFP signals were visualized without antibody staining).

Reagents used for determining gene expression patterns

Sema1-GFP (BL-60140, from Bloomington Drosophila Stock Center) is a Mi{MIC} insertion that results in an EGFP-tagged fusion protein. Plum-GFP is a protein fusion encoded by a genomic BAC transgene supplied by Dr. Oren Schuldiner [73]. CycB-GFP (BL-51568, from Bloomington Drosophila Stock Center) is a protein trap (GFP flanked by splice sites) transposon insertion into cycB that results in a GFP-tagged fusion protein. Santa maria-GAL4 (BL-24521, from Bloomington Drosophila Stock Center) is a P-element insertion of a transgene that includes promoter sequences from − 2030 to + 30 of the santa maria gene upstream of yeast GAL4 coding sequence. Rabbit anti-Dystroglycan antibody was supplied by Dr. Stefan Baumgartner [72]. All samples were processed for immunocytochemistry as described above for MARCM lineages, with GFP or Dystroglycan primary antibodies visualized as described, or RFP protein from santa maria-GAL4 > UAS-RFP visualized directly (without antibody staining), as in G-trace experiments.

Availability of data and materials

The datasets generated and/or analyzed during the current study are available in the repository, Gene Expression Omnibus (GEO) database, https://www.ncbi.nlm.nih.gov/geo (accession number GEO: GSE208674). Additionally, a large fraction of the scRNA-seq data, organized in informative formats, is presented in Additional files 4, 5, 8 and 9.

Abbreviations

FSC:: Follicle stem cell
EC:: Escort cell
r1 EC:: Region 1 escort cell
r2a EC:: Region 2a escort cell
FC:: Follicle cell
AP:: Anterior-posterior
DV:: Dorsoventral
scRNA:: Single-cell RNA
scRNA-seq:: Single-cell RNA sequencing
Wnt:: Wingless/Integrated
JAK-STAT:: Janus Kinase-Signal Transducer and Activator of Transcription
TF:: Terminal filament
GSC:: Germline stem cell
Fas3:: Fasciclin 3
FACS:: Fluorescence-activated cell sorting
UMI:: Unique molecular identifier
GEM:: Gel bead in emulsion
PCA:: Principal component analysis
MARCM:: Mosaic analysis with repressible cell marker
IGS:: Inner germarial sheath
Eya :: Eyes absent
Ilp6 :: Insulin-like peptide 6
croc :: Crocodile
wun 2 :: wunen 2
bnb :: Bangles and beads
Hf :: Helical factor
Netrin A :: NetA
Cre-loxP :: Causes recombination-locus of crossing-over (x), P1
PCNA :: Proliferating Cell Nuclear Antigen
Cas :: Castor
PI3K:: Phospho-inositide 3’ kinase
Hh:: Hedgehog
BMP:: Bone morphogenetic protein
EGFR:: Epidermal growth factor receptor
FGFR:: Fibroblast growth factor receptor
Sema 1a :: Semaphorin 1a
PlexA :: Plexin A
GFP:: Green fluorescent protein
RFP:: Red fluorescent protein
ECM:: Extracellular matrix
TGF-β:: Transforming growth factor β
CycB:: Cyclin B
t-SNE:: t-distributed stochastic neighbor embedding
PBS:: Phosphate-buffered saline

References

Clevers H, Watt FM. Defining adult stem cells by function, not by phenotype. Annu Rev Biochem. 2018;87:1015–27.
Article CAS PubMed Google Scholar
Kalderon D. Investigating adult stem cells through lineage analyses. Stem Cell Rev Rep. 2021;18(1):2–22.
Post Y, Clevers H. Defining adult stem cell function at its simplest: the ability to replace lost cells through mitosis. Cell Stem Cell. 2019;25(2):174–83.
Article CAS PubMed Google Scholar
Chatzeli L, Simons BD. Tracing the dynamics of stem cell fate. Cold Spring Harbor Perspect Biol. 2020;12(6):a036202.
Dray N, Than-Trong E, Bally-Cuif L. Neural stem cell pools in the vertebrate adult brain: Homeostasis from cell-autonomous decisions or community rules? BioEssays. 2021;43(3): e2000228.
Article PubMed Google Scholar
Goodell MA, Nguyen H, Shroyer N. Somatic stem cell heterogeneity: diversity in the blood, skin and intestinal stem cell compartments. Nat Rev Mol Cell Biol. 2015;16(5):299–309.
Article CAS PubMed PubMed Central Google Scholar
Jones PH. Stem cell fate in proliferating tissues: equal odds in a game of chance. Dev Cell. 2010;19(4):489–90.
Article CAS PubMed Google Scholar
Stine RR, Matunis EL. Stem cell competition: finding balance in the niche. Trends Cell Biol. 2013;23(8):357–64.
Article CAS PubMed PubMed Central Google Scholar
Reilein A, Melamed D, Tavare S, Kalderon D. Division-independent differentiation mandates proliferative competition among stem cells. Proc Natl Acad Sci U S A. 2018;115(14):E3182–91.
Article CAS PubMed PubMed Central Google Scholar
Ritsma L, Ellenbroek SI, Zomer A, Snippert HJ, de Sauvage FJ, Simons BD, et al. Intestinal crypt homeostasis revealed at single-stem-cell level by in vivo live imaging. Nature. 2014;507(7492):362–5.
Article CAS PubMed PubMed Central Google Scholar
Rompolas P, Mesa KR, Kawaguchi K, Park S, Gonzalez D, Brown S, et al. Spatiotemporal coordination of stem cell commitment during epidermal homeostasis. Science. 2016;352(6292):1471–4.
Article CAS PubMed PubMed Central Google Scholar
Greulich P, Simons BD. Dynamic heterogeneity as a strategy of stem cell self-renewal. Proc Natl Acad Sci U S A. 2016;113(27):7509–14.
Article CAS PubMed PubMed Central Google Scholar
Reilein A, Melamed D, Park KS, Berg A, Cimetta E, Tandon N, et al. Alternative direct stem cell derivatives defined by stem cell location and graded Wnt signalling. Nat Cell Biol. 2017;19(5):433–44.
Article CAS PubMed PubMed Central Google Scholar
Beumer J, Clevers H. Cell fate specification and differentiation in the adult mammalian intestine. Nat Rev Mol Cell Biol. 2020;22(1):39–53.
Donati G, Watt FM. Stem cell heterogeneity and plasticity in epithelia. Cell Stem Cell. 2015;16(5):465–76.
Article CAS PubMed Google Scholar
Greenspan LJ, de Cuevas M, Le KH, Viveiros JM, Matunis EL. Activation of the EGFR/MAPK pathway drives transdifferentiation of quiescent niche cells to stem cells in the Drosophila testis niche. Elife. 2022;11:e70810.
Melamed D, Kalderon D. Opposing JAK-STAT and Wnt signaling gradients define a stem cell domain by regulating differentiation at two borders. Elife. 2020;9:e61204.
Rust K, Nystul T. Signal transduction in the early Drosophila follicle stem cell lineage. Curr Opin Insect Sci. 2020;37:39–48.
Article PubMed PubMed Central Google Scholar
Bastock R, St JD. Drosophila oogenesis. Curr Biol. 2008;18(23):R1082–7.
Article CAS PubMed Google Scholar
Eliazer S, Buszczak M. Finding a niche: studies from the Drosophila ovary. Stem Cell Res Ther. 2011;2(6):45.
Article PubMed PubMed Central Google Scholar
Spradling A, Fuller MT, Braun RE, Yoshida S. Germline stem cells. Cold Spring Harb Perspect Biol. 2011;3(11): a002642.
Article PubMed PubMed Central Google Scholar
Decotto E, Spradling AC. The Drosophila ovarian and testis stem cell niches: similar somatic stem cells and signals. Dev Cell. 2005;9(4):501–10.
Article CAS PubMed Google Scholar
Kirilly D, Wang S, Xie T. Self-maintained escort cells form a germline stem cell differentiation niche. Development. 2011;138(23):5087–97.
Article CAS PubMed PubMed Central Google Scholar
Dai W, Peterson A, Kenney T, Burrous H, Montell DJ. Quantitative microscopy of the Drosophila ovary shows multiple niche signals specify progenitor cell fate. Nat Commun. 2017;8(1):1244.
Article PubMed PubMed Central Google Scholar
Margolis J, Spradling A. Identification and behavior of epithelial stem cells in the Drosophila ovary. Development. 1995;121(11):3797–807.
Article CAS PubMed Google Scholar
Tworoger M, Larkin MK, Bryant Z, Ruohola-Baker H. Mosaic analysis in the drosophila ovary reveals a common hedgehog-inducible precursor stage for stalk and polar cells. Genetics. 1999;151(2):739–48.
Article CAS PubMed PubMed Central Google Scholar
Duhart JC, Parsons TT, Raftery LA. The repertoire of epithelial morphogenesis on display: Progressive elaboration of Drosophila egg structure. Mech Dev. 2017;148:18–39.
Article CAS PubMed Google Scholar
Riechmann V, Ephrussi A. Axis formation during Drosophila oogenesis. Curr Opin Genet Dev. 2001;11(4):374–83.
Article CAS PubMed Google Scholar
Zhang Y, Kalderon D. Hedgehog acts as a somatic stem cell factor in the Drosophila ovary. Nature. 2001;410(6828):599–604.
Article CAS PubMed Google Scholar
Melamed D, Choi A, Reilein A, Kalderon D. Spatial regulation of Drosophila ovarian Follicle Stem Cell division rates and cell cycle transitions. bioRxiv. 2022. https://doi.org/10.1101/2022.06.22.497017.
Jevitt A, Chatterjee D, Xie G, Wang XF, Otwell T, Huang YC, et al. A single-cell atlas of adult Drosophila ovary identifies transcriptional programs and somatic cell lineage regulating oogenesis. PLoS Biol. 2020;18(4): e3000538.
Article PubMed PubMed Central Google Scholar
Slaidina M, Gupta S, Banisch TU, Lehmann R. A single-cell atlas reveals unanticipated cell type complexity in Drosophila ovaries. Genome Res. 2021;31(10):1938–51.
Article PubMed PubMed Central Google Scholar
Forbes AJ, Spradling AC, Ingham PW, Lin H. The role of segment polarity genes during early oogenesis in Drosophila. Development. 1996;122(10):3283–94.
Article CAS PubMed Google Scholar
Godt D, Laski FA. Mechanisms of cell rearrangement and cell recruitment in Drosophila ovary morphogenesis and the requirement of bric a brac. Development. 1995;121(1):173–87.
Article CAS PubMed Google Scholar
Xie T, Spradling AC. decapentaplegic is essential for the maintenance and division of germline stem cells in the Drosophila ovary. Cell. 1998;94(2):251–60.
Article CAS PubMed Google Scholar
Allbee AW, Rincon-Limas DE, Biteau B. Lmx1a is required for the development of the ovarian stem cell niche in Drosophila. Development. 2018;145(8):dev163394.
Buszczak M, Paterno S, Lighthouse D, Bachman J, Planck J, Owen S, et al. The carnegie protein trap library: a versatile tool for Drosophila developmental studies. Genetics. 2007;175(3):1505–31.
Article CAS PubMed PubMed Central Google Scholar
Vied C, Reilein A, Field NS, Kalderon D. Regulation of stem cells by intersecting gradients of long-range niche signals. Dev Cell. 2012;23(4):836–48.
Article CAS PubMed PubMed Central Google Scholar
Waghmare I, Page-McCaw A. Wnt signaling in stem cell maintenance and differentiation in the Drosophila Germarium. Genes (Basel). 2018;9(3):127.
Tu R, Duan B, Song X, Chen S, Scott A, Hall K, et al. Multiple niche compartments orchestrate stepwise germline stem cell progeny differentiation. Curr Biol. 2021;31(4):827-39 e3.
Article CAS PubMed Google Scholar
McGregor JR, Xi R, Harrison DA. JAK signaling is somatically required for follicle cell differentiation in Drosophila. Development. 2002;129(3):705–17.
Article CAS PubMed Google Scholar
Reilein A, Kogan HV, Misner R, Park KS, Kalderon D. Adult stem cells and niche cells segregate gradually from common precursors that build the adult Drosophila ovary during pupal development. Elife. 2021;10:e69749.
Wang X, Page-McCaw A. Wnt6 maintains anterior escort cells as an integral component of the germline stem cell niche. Development. 2018;145(3):dev158527.
Kretzschmar K, Watt FM. Lineage tracing. Cell. 2012;148(1–2):33–45.
Article CAS PubMed Google Scholar
Evans CJ, Olson JM, Ngo KT, Kim E, Lee NE, Kuoy E, et al. G-TRACE: rapid Gal4-based cell lineage analysis in Drosophila. Nat Methods. 2009;6(8):603–5.
Article CAS PubMed PubMed Central Google Scholar
Rust K, Byrnes LE, Yu KS, Park JS, Sneddon JB, Tward AD, et al. A single-cell atlas and lineage analysis of the adult Drosophila ovary. Nat Commun. 2020;11(1):5628.
Article CAS PubMed PubMed Central Google Scholar
Shi J, Jin Z, Yu Y, Zhang Y, Yang F, Huang H, et al. A progressive somatic cell niche regulates germline cyst differentiation in the drosophila ovary. Curr Biol. 2021;31(4):840-52 e5.
Article CAS PubMed Google Scholar
Chung S, Le TP, Vishwakarma V, Cheng YL, Andrew DJ. Isoform-specific roles of the Drosophila filamin-type protein Jitterbug (Jbug) during development. Genetics. 2021;219(2):iyab100.
Wang X, Page-McCaw A. A matrix metalloproteinase mediates long-distance attenuation of stem cell proliferation. J Cell Biol. 2014;206(7):923–36.
Article CAS PubMed PubMed Central Google Scholar
Sun J, Deng WM. Notch-dependent downregulation of the homeodomain gene cut is required for the mitotic cycle/endocycle switch and cell differentiation in Drosophila follicle cells. Development. 2005;132(19):4299–308.
Article CAS PubMed Google Scholar
Wang ZA, Huang J, Kalderon D. Drosophila follicle stem cells are regulated by proliferation and niche adhesion as well as mitochondria and ROS. Nat Commun. 2012;3:769.
Article PubMed Google Scholar
Song X, Xie T. DE-cadherin-mediated cell adhesion is essential for maintaining somatic stem cells in the Drosophila ovary. Proc Natl Acad Sci USA. 2002;99(23):14813–8.
Article CAS PubMed PubMed Central Google Scholar
Hartman TR, Zinshteyn D, Schofield HK, Nicolas E, Okada A, O’Reilly AM. Drosophila Boi limits Hedgehog levels to suppress follicle stem cell proliferation. J Cell Biol. 2010;191(5):943–52.
Article CAS PubMed PubMed Central Google Scholar
Huang J, Kalderon D. Coupling of Hedgehog and Hippo pathways promotes stem cell maintenance by stimulating proliferation. J Cell Biol. 2014;205(3):325–38.
Article CAS PubMed PubMed Central Google Scholar
O’Reilly AM, Lee HH, Simon MA. Integrins control the positioning and proliferation of follicle stem cells in the Drosophila ovary. J Cell Biol. 2008;182(4):801–15.
Article CAS PubMed PubMed Central Google Scholar
Herrera SC, Sainz de la Maza D, Grmai L, Margolis S, Plessel R, Burel M, et al. Proliferative stem cells maintain quiescence of their niche by secreting the Activin inhibitor Follistatin. Dev Cell. 2021;56(16):2284–94.
Upadhyay M, Kuna M, Tudor S, Martino Cortez Y, Rangan P. A switch in the mode of Wnt signaling orchestrates the formation of germline stem cell differentiation niche in Drosophila. PLoS Genet. 2018;14(1): e1007154.
Article PubMed PubMed Central Google Scholar
Harris RE, Ashe HL. Cease and desist: modulating short-range Dpp signalling in the stem-cell niche. EMBO Rep. 2011;12(6):519–26.
Article CAS PubMed PubMed Central Google Scholar
Ingham PW. Hedgehog signaling. Curr Top Dev Biol. 2022;149:1–58.
Article CAS PubMed Google Scholar
Forbes AJ, Lin H, Ingham PW, Spradling AC. hedgehog is required for the proliferation and specification of ovarian somatic cells prior to egg chamber formation in Drosophila. Development. 1996;122(4):1125–35.
Article CAS PubMed Google Scholar
Sahai-Hernandez P, Nystul TG. A dynamic population of stromal cells contributes to the follicle stem cell niche in the Drosophila ovary. Development. 2013;140(22):4490–8.
Article CAS PubMed PubMed Central Google Scholar
Kakugawa S, Langton PF, Zebisch M, Howell S, Chang TH, Liu Y, et al. Notum deacylates Wnt proteins to suppress signalling activity. Nature. 2015;519(7542):187–92.
Article CAS PubMed PubMed Central Google Scholar
Zeng W, Wharton KA Jr, Mack JA, Wang K, Gadbaw M, Suyama K, et al. naked cuticle encodes an inducible antagonist of Wnt signalling. Nature. 2000;403(6771):789–95.
Article CAS PubMed Google Scholar
Matsuoka S, Gupta S, Suzuki E, Hiromi Y, Asaoka M. gone early, a novel germline factor, ensures the proper size of the stem cell precursor pool in the Drosophila ovary. PLoS ONE. 2014;9(11): e113423.
Article PubMed PubMed Central Google Scholar
Sapir A, Schweitzer R, Shilo BZ. Sequential activation of the EGF receptor pathway during Drosophila oogenesis establishes the dorsoventral axis. Development. 1998;125(2):191–200.
Article CAS PubMed Google Scholar
Baeg GH, Lin X, Khare N, Baumgartner S, Perrimon N. Heparan sulfate proteoglycans are critical for the organization of the extracellular distribution of Wingless. Development. 2001;128(1):87–94.
Article CAS PubMed Google Scholar
Avanesov A, Blair SS. The Drosophila WIF1 homolog Shifted maintains glypican-independent Hedgehog signaling and interacts with the Hedgehog co-receptors Ihog and Boi. Development. 2013;140(1):107–16.
Article CAS PubMed PubMed Central Google Scholar
Han C, Belenkaya TY, Wang B, Lin X. Drosophila glypicans control the cell-to-cell movement of Hedgehog by a dynamin-independent process. Development. 2004;131(3):601–11.
Article CAS PubMed Google Scholar
Yu L, Zhou Y, Cheng S, Rao Y. Plexin a-semaphorin-1a reverse signaling regulates photoreceptor axon guidance in Drosophila. J Neurosci. 2010;30(36):12151–6.
Article CAS PubMed PubMed Central Google Scholar
Stedden CG, Menegas W, Zajac AL, Williams AM, Cheng S, Ozkan E, et al. Planar-polarized semaphorin-5c and plexin A promote the collective migration of epithelial cells in drosophila. Curr Biol. 2019;29(6):908-20 e6.
Article CAS PubMed Google Scholar
Cerqueira Campos F, Dennis C, Alegot H, Fritsch C, Isabella A, Pouchin P, et al. Oriented basement membrane fibrils provide a memory for F-actin planar polarization via the Dystrophin-Dystroglycan complex during tissue elongation. Development. 2020;147(7):dev186957.
Schneider M, Khalil AA, Poulton J, Castillejo-Lopez C, Egger-Adam D, Wodarz A, et al. Perlecan and Dystroglycan act at the basal side of the Drosophila follicular epithelium to maintain epithelial organization. Development. 2006;133(19):3805–15.
Article CAS PubMed Google Scholar
Yu XM, Gutman I, Mosca TJ, Iram T, Ozkan E, Garcia KC, et al. Plum, an immunoglobulin superfamily protein, regulates axon pruning by facilitating TGF-beta signaling. Neuron. 2013;78(3):456–68.
Article CAS PubMed PubMed Central Google Scholar
Zielke N, Korzelius J, van Straaten M, Bender K, Schuhknecht GF, Dutta D, et al. Fly-FUCCI: a versatile tool for studying cell proliferation in complex tissues. Cell Rep. 2014;7(2):588–98.
Article CAS PubMed Google Scholar
Wang X, Wang T, Jiao Y, von Lintig J, Montell C. Requirement for an enzymatic visual cycle in Drosophila. Curr Biol. 2010;20(2):93–102.
Article CAS PubMed Google Scholar
Fadiga J, Nystul TG. The follicle epithelium in the Drosophila ovary is maintained by a small number of stem cells. Elife. 2019;8:e49050.
Kalderon D, Melamed D, Reilein A. Follicle Stem Cells (FSCs) in the Drosophila ovary; a critique of published studies defining the number, location and behavior of FSCs. bioRxiv. 2020:2020.06.25.171579. https://doi.org/10.1101/2020.06.25.171579.
Kalderon D, Melamed D, Reilein A. The number of follicle stem cells in a Drosophila ovariole. bioRxiv. 2021:2021.10.25.465475. https://doi.org/10.1101/2021.10.25.465475.
Zheng GX, Terry JM, Belgrader P, Ryvkin P, Bent ZW, Wilson R, et al. Massively parallel digital transcriptional profiling of single cells. Nat Commun. 2017;8:14049.
Article CAS PubMed PubMed Central Google Scholar
Hao Y, Hao S, Andersen-Nissen E, Mauck WM 3rd, Zheng S, Butler A, et al. Integrated analysis of multimodal single-cell data. Cell. 2021;184(13):3573-87 e29.
Article CAS PubMed PubMed Central Google Scholar

Download references

Acknowledgements

We thank John Lis (Cornell University) and the Bloomington stock center for provision of genetic reagents, the Developmental Studies Hybridoma Bank (DSHB) for antibodies, FlyBase as an information resource, David Melamed for preparation of Fig. 1, and the confocal microscope resource provided by the Dept. of Biological Sciences, Columbia University.

Funding

This work was supported by the National Science Foundation of China 32172467 to JH, National Key Research and Development Program of China 2019YFA0111500 to XL, NIH RO1 GM079351 to DK, and a Columbia University Department of Biological Sciences SURF undergraduate research fellowship to XT.

Author information

Zhi Dong and Lan Pang contributed equally to this work.

Authors and Affiliations

Institute of Insect Sciences, Ministry of Agriculture Key Lab of Molecular Biology of Crop Pathogens and Insect Pests, College of Agriculture and Biotechnology, Zhejiang University, Hangzhou, 310058, China
Zhi Dong, Lan Pang, Zhiguo Liu, Yifeng Sheng & Jianhua Huang
Department of Hepatic Surgery and Liver Transplantation Center of the Third Affiliated Hospital, Organ Transplantation Institute, Sun Yat-Sen University, Guangzhou, 510630, Guangdong, China
Xiaoping Li
Department of Biological Sciences, Columbia University, New York, NY, USA
Xavier Thibault, Amy Reilein & Daniel Kalderon

Authors

Zhi Dong
View author publications
You can also search for this author in PubMed Google Scholar
Lan Pang
View author publications
You can also search for this author in PubMed Google Scholar
Zhiguo Liu
View author publications
You can also search for this author in PubMed Google Scholar
Yifeng Sheng
View author publications
You can also search for this author in PubMed Google Scholar
Xiaoping Li
View author publications
You can also search for this author in PubMed Google Scholar
Xavier Thibault
View author publications
You can also search for this author in PubMed Google Scholar
Amy Reilein
View author publications
You can also search for this author in PubMed Google Scholar
Daniel Kalderon
View author publications
You can also search for this author in PubMed Google Scholar
Jianhua Huang
View author publications
You can also search for this author in PubMed Google Scholar

Contributions

Conceptualization, J.H. and D.K., Formal Analysis, Z.D., J.H. and D.K., Funding Acquisition, J. H. and D.K., Investigation and Methodology, L.P., Z.L., Y.S. (scRNA-seq experiment), Z.D., X.L. (bioinformatics), X.T., A.R. D.K. (G-trace experiments), Supervision, J.H., A.R. and D.K., Validation, Z.D., J.H., A.R., D.K., Visualization, X.T, A.R., Writing-Original Draft, Z.D., J.H., D.K., Writing-Review & Editing, Z.D., J.H., D.K.. All authors read and approved the final manuscript.

Corresponding authors

Correspondence to Daniel Kalderon or Jianhua Huang.

Ethics declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

Not applicable.

Competing interests

The authors declare that they have no competing interests.

Additional information

Publisher's Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Supplementary Information

Additional file 1: Figure S1.

Identity markers for peripheral groups in initial t-SNE clusters. Related to Figure 2.t-SNE plots of complete data setshowing color-coded relative expression levels of genes characteristic ofTF or cap cells,stalk or pre-stalk cellsorgermline cells.

Additional File 2: Table 1 for Fig. 2C.

Cell barcodes. Cell barcodes for each dot on Fig. 2C.

Additional File 3: Table 2 for Fig. 2E.

Expression characteristics for key cluster genes. Average expression values and percentage of cells with detected expression underlying the dot size and coloring in Fig. 2E.

Additional File 4: Table 3.

Ranked list of genes characteristic of each cluster. Gene order for each cluster is determined by the ratio of expression in the characteristic cluster relative to average expression in all other cells. This enrichmentvalue for cluster “X” is calculated as E =/and denoted “avg_log2FC”. Also shown are the raw p values for significant differences in expression, adjusted p values after Bonferroni correction for multiple samples, the percentage of cells where the gene is detected in the characteristic clusterand in all other cells, as well as the normalized UMI counts for each cluster. Clusterscorrespond to Fig. 2C; each cluster is colored differently. This spreadsheet lists genes in each cluster down to an E value of 0.4. These genes and others with lower E values are listed in Additional File 5: Table 4.

Additional File 5: Table 4.

Genes characteristic of each cluster. Genes characteristic of each cluster are listed together with information characteristic of Additional File 4: Table 1 in columns B-G and information characteristic of Additional File 8: Table 7 or Additional File 9: Table 8 in remaining columns. The threshold for inclusion in a cluster here is extremely lowcompared to Additional File 4: Table 3, to include 7547 genes.

Additional File 6: Table 5 for Fig. 6A.

PC1 values. All genes listed by descending values for PC1.

Additional File 7: Table 6 for Fig. 6B.

PC2 values. All genes listed by descending values for PC2.

Additional File 8: Table 7.

Genes with increasing expression from ECs to FSCs and/or FSCs to FCs. Average normalized expression value for each cluster is listed in columns I-N for each groupand weighted values used to calculate expression among ECsand FCs. Fractional changes relative to the larger number were calculated for transitions from ECs to FSCs, FSCs to FCs, group 5 to group 1and group 1 to group 4. The resulting values in those four columns were classified into quartiles, with the first quartile colored dark brown for an increase or dark blue for a decrease, the second quartile colored light brown for an increase or light blue for a decrease, and the third quartile colored very light brown for an increase or very light blue for a decrease. Gene order was determined according to quartiles, first prioritizing genes that increased for both transitions, then genes increased only for the EC to FSC transition, listing first quartile genes first in each group. Horizontal green lines separate different categories and gene numbering begins at 1 for each category. Additional File 5: Table 4 lists all genes with latter columns in the same format as in this spreadsheet, regardless of the magnitude of changes from ECs to FSCs or FSCs to FCs.

Additional File 9: Table 8.

Genes with decreasing expression from ECs to FSCs and/or FSCs to FCs. Average normalized expression value for each cluster is listed in columns I-N for each groupand weighted values used to calculate expression among ECsand FCs. Fractional changes relative to the larger number were calculated for transitions from ECs to FSCs, FSCs to FCs, group 5 to group 1and group 1 to group 4. The resulting values in those four columns were classified into quartiles, with the first quartile colored dark brown for an increase or dark blue for a decrease, the second quartile colored light brown for an increase or light blue for a decrease, and the third quartile colored very light brown for an increase or very light blue for a decrease. Gene order was determined according to quartiles, first prioritizing genes that decreased for both transitions, then genes decreased only for the EC to FSC transition, listing first quartile genes first in each group. Horizontal green lines separate different categories and gene numbering begins at 1 for each category. Additional File 5: Table 4 lists all genes with latter columns in the same format as in this spreadsheet, regardless of the magnitude of changes from ECs to FSCs or FSCs to FCs.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/. The Creative Commons Public Domain Dedication waiver (http://creativecommons.org/publicdomain/zero/1.0/) applies to the data made available in this article, unless otherwise stated in a credit line to the data.

Reprints and permissions

About this article

Cite this article

Dong, Z., Pang, L., Liu, Z. et al. Single-cell expression profile of Drosophila ovarian follicle stem cells illuminates spatial differentiation in the germarium. BMC Biol 21, 143 (2023). https://doi.org/10.1186/s12915-023-01636-9

Download citation

Received: 22 July 2022
Accepted: 26 May 2023
Published: 20 June 2023
DOI: https://doi.org/10.1186/s12915-023-01636-9

Single-cell expression profile of Drosophila ovarian follicle stem cells illuminates spatial differentiation in the germarium

Abstract

Background

Results

Conclusions

Background

Results

Location of FSCs

scRNA-seq investigation of anterior somatic germarial cells

Cluster boundaries and cell identities

Prospective FSC labeling for lineage analyses

Patterns of gene expression

Potential regulators of FSC behavior

Independent assessment of graded expression patterns suggested by scRNA-seq

Discussion

Conclusions

Methods

Single-cell suspension preparation and fluorescence-activated cell sorting (FACS)

Single-cell RNA sequencing (scRNA-seq)

Raw data preprocessing

Single-cell RNA-seq data analysis

MARCM and multicolor FSC lineages

Protocol for identifying cell locations relative to landmark of anterior border of strong Fas3 staining

G-trace experiments

Reagents used for determining gene expression patterns

Availability of data and materials

Abbreviations

References

Acknowledgements

Funding

Author information

Authors and Affiliations

Contributions

Corresponding authors

Ethics declarations

Ethics approval and consent to participate

Consent for publication

Competing interests

Additional information

Publisher's Note

Supplementary Information

Rights and permissions

About this article

Cite this article

Share this article

Keywords

BMC Biology

Contact us