Sperm-contributed centrioles segregate stochastically into blastomeres of 4-cell stage Caenorhabditis elegans embryos

Abstract Whereas both sperm and egg contribute nuclear genetic material to the zygote in metazoan organisms, the inheritance of other cellular constituents is unequal between the 2 gametes. Thus, 2 copies of the centriole are contributed solely by the sperm to the zygote in most species. Centrioles can have a stereotyped distribution in some asymmetric divisions, but whether sperm-contributed centrioles are distributed in a stereotyped manner in the resulting embryo is not known. Here, we address this question in Caenorhabditis elegans using marked mating experiments, whereby the presence of the 2 sperm-contributed centrioles is monitored in the embryo using the stable centriolar component SAS-4::GFP, as well as GFP::SAS-7. Our analysis demonstrates that the distribution of sperm-contributed centrioles is stochastic in 4-cell stage embryos. Moreover, using sperm from zyg-1 mutant males that harbor a single centriole, we show that the older sperm-contributed centriole is likewise distributed stochastically in the resulting embryo. Overall, we conclude that, in contrast to the situation during some asymmetric cell divisions, centrioles contributed by the male germ line are distributed stochastically in embryos of C. elegans.


Introduction
Whereas the genetic material is contributed in an equal manner to the zygote by the 2 parental gametes of metazoan organisms, this is not the case for the remainder of the cellular material. In most species, the egg contributes the bulk of the messenger RNAs, proteins, and cytoplasmic constituents to the zygote. The same holds true for mitochondria, which are inherited and retained from maternal stores by the developing embryo. Conversely, the sperm is the sole contributor of centrioles, which are absent from the egg (reviewed by Schatten 1994; Delattre and Gönczy 2004). Such differential contribution of centrioles is critical for endowing the zygote with strictly 2 centrioles at the onset of life. Whether the 2 sperm-contributed centrioles are segregated thereafter to specific cells of the developing embryo is not known.
Centrioles are small microtubule-based organelles that are critical for fundamental cellular processes across eukaryotes (reviewed by Bornens 2012; Winey and O'Toole 2014). Centrioles are key notably in their capacity as basal bodies that seed the formation of the axoneme of cilia and flagella in specific cell types, including sperm cells of most species. Moreover, in animal systems, centrioles recruit the pericentriolar material (PCM), thus forming the centrosome, which acts as an important microtubule-organizing center. Through this role, centrioles are important for fundamental cellular processes such as polarity and division.
Centrosomes have an inherent cell generational asymmetry: most proliferating cells are born with 2 centrioles, a so-called mother centriole that has been generated at least 1 cell generation prior and a so-called daughter centriole that has been generated during the previous cell cycle (reviewed by Nigg and Stearns 2011). Toward the onset of S phase, both mother and daughter centrioles seed the formation of a procentriole in their vicinity, yielding 2 centrosomes, each containing a centriole/procentriole pair, which direct bipolar spindle assembly during mitosis.
Interestingly, centriole inheritance is stereotyped in some cases of asymmetric cell division (reviewed by Yamashita 2009). For instance, in germ line stem cells of the Drosophila testis, the centrosome harboring the older centriole, which is referred to as the older centrosome, is inherited systematically by the cell that retains the stem cell fate (Yamashita et al. 2007). Perturbing such stereotyped inheritance through mutation of the PCM protein centrosomin randomizes centrosome distribution and results in declined stem cell function (Yamashita et al. 2007). Conversely, the older centrosome is inherited systematically by the ganglion mother cell during asymmetric division of Drosophila neuroblasts, whereas stem-like neuroblasts retain the younger centrosome (Conduit and Raff 2010;Januschke et al. 2011). Such stereotyped asymmetric centriole inheritance is not limited to Drosophila. Thus, radial glial progenitors in the ventricular zone of the developing mouse cortex also exhibit stereotyped centrosome inheritance (Wang et al. 2009). In this case, the mother centriole is inherited systematically by the radial glial progenitor stem cell during asymmetric division, reminiscent of the situation in the fly testis. Moreover, depletion of the mother centriole-specific protein Ninein, which is essential for mother centriole anchoring to the plasma membrane, randomizes centrosome distribution in this setting. Importantly, this is accompanied by premature depletion of radial glial progenitor from the ventricular zone (Wang et al. 2009).
In contrast to the wealth of information regarding the inheritance pattern of centrosomes and its importance in stem cell systems, little is known regarding the distribution of the 2 sperm-contributed centrioles in the resulting developing embryos. Conceivably, such inheritance could be stereotyped as well and thereby potentially endow specific embryonic blastomeres with paternally derived components. This paucity of information stems notably from the fact that centrioles assembled in the zygote using maternal components can be difficult to distinguish from those contributed by sperm. This experimental limitation can be circumvented with so-called marked mating experiments, which have been deployed in Caenorhabditis elegans and which are conceptually analogous to experiments in mammalian cells that revealed the mode of centriole duplication (Kochanski and Borisy 1990). In marked mating experiments in the worm (Kirkham et al. 2003;Leidel and Gönczy 2003;Balestra et al. 2015), a stable centriolar component is labeled in sperm with a fluorescent protein such as GFP. Upon mating with an egg devoid of this GFP-tagged centriolar component, the 2 sperm-contributed centrioles stably harbor GFP, whereas all centrioles generated in the zygote do not. Therefore, this experimental setting offers an optimal means to monitor the fate of paternally contributed centrioles in early embryos.

Experimental design
We set out to use marked mating experiments to test whether sperm-contributed centrioles are distributed in early C. elegans embryos in a stereotyped fashion, or instead stochastically, focusing our analysis on the 4-cell stage. The experimental strategy is summarized schematically in Fig. 1. Hermaphrodites expressing TagRFP-T::SAS-7 and homozygous for fem-1(hc17ts) are raised at 25°C, which results in them lacking sperm. Such feminized hermaphrodites are then mated to males expressing endogenously tagged SAS-4::GFP or GFP::SAS-7 (Fig. 1a). As a result, the 2 spermcontributed centrioles are marked with GFP (Fig. 1b), and each such centriole seeds the formation of 1 procentriole in its vicinity during the first cell cycle in the zygote using maternal-contributing components, since zygotic transcription begins only later. The 2 resulting centrosomes, each with a centriole/procentriole pair, constitute the 2 poles of the mitotic spindle at the end of the first cell cycle (Fig. 1c, left). Each resulting daughter cell, termed AB and P 1 , respectively, will necessarily inherit 1 of the 2 sperm-contributed centrioles (Fig. 1c, right). At the onset of the second cell cycle, each of the 4 centrioles that are present at that time seeds the formation of 1 procentriole in its vicinity (Fig. 1c, right). As illustrated in Fig. 1d, this could in principle yield 4 distinct distributions of spermcontributed centrioles at the following cell cycle, in 4-cell stage embryos: ABa and P 2 , ABa and EMS, ABp and P 2 , or ABp and EMS.

Stochastic distribution of sperm-contributed centrioles at the 4-cell stage
We deployed the above experimental strategy using primarily SAS-4::GFP because prior work established that this evolutionarily conserved centriolar protein does not undergo significant exchange once incorporated in the centriole (Kirkham et al. 2003;Leidel and Gönczy 2003;Balestra et al. 2015). As illustrated in Fig. 2a and Supplementary Fig. 1, the TagRFP-T::SAS-7 signal serves to identify the location of centrioles in each cell, whereas the SAS-4::GFP signal is monitored to determine the distribution of sperm-contributed centrioles. Importantly, this analysis uncovered that the fraction of 4-cell embryos harboring a sperm-contributed centriole in ABa is indistinguishable from that harboring a sperm-contributed centriole in ABp ( Tables  1 and 2). Moreover, there is no stereotyped pair-wise arrangement of cells harboring sperm-contributed centrioles; instead, the fraction of embryos with sperm-contributed centrioles in ABa and P 2 is indistinguishable from that in ABa and EMS, ABp and P 2 , or ABp and EMS ( To test whether this outcome is specific to SAS-4::GFP, we conducted analogous experiments using GFP::SAS-7, which exhibits a strong signal in sperm centrioles, making it a potentially suitable candidate for marked mating experiments. We found, however, that part of the GFP::SAS-7 signal present in sperm centrioles is lost shortly after fertilization, such that we relied primarily on immunofluorescence analysis in this case to more reliably detect sperm-contributed centriolar GFP::SAS-7 in 4-cell stage embryos. We used antibodies against the centriolar marker IFA to identify all centrioles, as well as against GFP to detect specifically sperm-contributed centrioles  Tables 1 and 2). Taken together, these experiments lead us to conclude that paternally contributed centrioles are distributed in a stochastic fashion in 4-cell stage C. elegans embryos.

The older sperm-contributed centriole can be present in any blastomere at the 4-cell stage
We next set out to address whether there might be a preferential inheritance of the older vs the younger sperm-contributed centriole in the resulting embryos. However, no marker is known in C. elegans that would distinguish the older centriole from the younger one. Proteins such as Ninein that are present specifically at the mother centriole in other systems are absent from the worm genome, and the 2 sperm-contributed centrioles appear similar by electron microscopy (Wolf et al. 1978). To bypass this limitation, we designed a genetic strategy relying on the  temperature-sensitive mutant allele zyg-1(b1ts) (Wood et al. 1980). ZYG-1 is a kinase essential for procentriole formation in C. elegans and a relative of the Polo-like-kinase PLK4 that exerts a similar function in other metazoan organisms (O'Connell et al. 2001).
The zyg-1(b1ts) mutant allele is endowed with normal function at the permissive temperature of 15°C but exhibits a strong and perhaps complete loss of function phenotype at the restrictive temperature of 25°C (Wood et al. 1980;O'Connell et al. 2001).
In this modified marked mating experiments, we shifted zyg-1(b1ts) mutant males from 15°C to 25°C during the third larval instar stage. This enables assembly of procentrioles in the proliferating mitotic zone of the gonad prior to the temperature shift but prevents their formation thereafter. As a result, most sperm cells derived from such animals contain a single centriole, which corresponds to the older one in the wild type, whereas the centriole that should have been generated next is missing (O'Connell et al. 2001). During the first cell cycle, this single sperm-contributed centriole seeds the formation of a procentriole in its vicinity. Because this centriole/procentriole pair remains joined until the end of the first cell cycle, a monopolar spindle assembles, leading to aberrant chromosome segregation and cleavage failure (Fig. 3a, left-most). In the second cell cycle, each centriole seeds the formation of a procentriole, leading to bipolar spindle assembly (Fig. 3a, second panel), and the generation of AB-like and P 1 -like cells in the third cell cycle (Fig. 3a, third panel). These 2 blastomeres then divide asynchronously, with the AB-like cell undergoing mitosis before the P 1 -like cell, as is the case for AB and P 1 in the second cell cycle of wild-type embryos, ultimately yielding embryos that harbor AB-like, ABp-like, EMS-like, and P 2 -like blastomeres (Fig. 3a, right-most).
Using live imaging of the entire sequence of events to ensure proper staging of embryos ( Fig. 3b and c; Supplementary Video 1), we addressed whether the single centriole contributed by sperm from zyg-1(b1ts) mutant males and harboring SAS-4::GFP segregates strictly to AB-derived or P 1 -derived cells, scoring blastomeres at the 4-cell stage. As shown in Fig. 3d and e, we found this not to be the case, as this single marked centriole can be present in either ABa-like, ABp-like, EMS-like, or P 2 -like blastomere (Supplementary Table 1). Although we cannot formally exclude that the aberrant divisions in this experimental settings might alter the distribution of the sole centriole, our findings indicate that the older sperm-contributed centriole is also distributed in a stochastic fashion in 4-cell stage embryos.

Discussion
Our findings demonstrate that, in contrast to the stereotypy of centriole distribution during some asymmetric divisions in flies and mammals, sperm-contributed centrioles are distributed in a stochastic manner in embryos of C. elegans. It could have been envisaged that sperm-contributed centrioles are excluded from certain cells at the 4-cell stage, for instance the germ line precursor P 2 , thus ensuring that only newly made centrioles are transmitted in this lineage. Our work demonstrates that this is not the case. Therefore, although sperm-contributed centrioles persist for several hours during C. elegans embryogenesis (Balestra et al. 2015), . 3. The older sperm-contributed centriole can be present in any blastomere at the 4-cell stage. a) Schematic of sequence of events in embryos following fertilization of fem-1(hc17ts) hermaphrodites expressing TagRFP-T::SAS-7, raised at 25°C and thus lacking sperm, by zyg-1(b1ts) mutant males shifted to 25°C during spermatogenesis, leading to sperm with only 1 centriole marked by SAS-4::GFP. During the first cell cycle, this single centriole seeds the formation of a procentriole in its vicinity, but the first division is monopolar, leading to aberrant chromosome segregation and cleavage failure (left).
In the second cell cycle, each centriole seeds the formation of a procentriole, such that a bipolar spindle assembles during mitosis (second panel), generating AB-like and P 1 -like cells in the third cell cycle (third panel). These cells divide asynchronously, with the AB-like cell going first, yielding embryos at the fourth cell cycle that harbor AB-like, ABp-like, EMS-like, and P 2 -like blastomeres (right). Color code as in Fig. 1a. Note that inheritance of the sole paternal centriole by ABa-like is represented here. b) DIC images from a time-lapse recording corresponding to the stages schematized in ( the present findings establish that they cannot endow specific blastomeres in a stereotyped fashion with potential transgenerational biological information. Regardless, it will be interesting to address whether sperm-derived centrioles exhibit preferential inheritance at later stages of development. In this context, we note that experiments with Dendra2::SAS-4 showed that the older centriole marked following photo-conversion can also segregate to either daughter cell of ABprpppaa and ABprpppap (Erpf and Mikeladze-Dvali 2020), indicating that stochastic inheritance in the worm is not restricted to sperm-contributed centrioles. It will also be interesting to investigate whether the random distribution of sperm-contributed centrioles uncovered here is characteristic of worms, especially considering that nematode sperm is not flagellated, in contrast to the situation in most other metazoan species. Moreover, whereas the 2 C. elegans spermcontributed centrioles are analogous at the ultrastructural level (Wolf et al. 1978), they are different from one another in other species, including Drosophila and man (Fawcett 1975;Blachon et al. 2009). Therefore, it is plausible that in those cases, such ultrastructural differences translate into stereotyped inheritance of sperm-contributed centrioles in the resulting embryos. Alternatively, the stochastic distribution revealed in our study might reflect a more general feature of the centriole organelle as it stands prior to fertilization.

Live imaging and analysis
Embryos were dissected from the uterus of hermaphrodites in a watch glass containing M9 and transferred using a mouth pipette onto a 2% agarose pad, which was overlaid gently with an 18 × 18 mm coverslip. Time-lapse DIC and dual color fluorescent microscopy imaging was performed at ∼21°C on a Zeiss Axioplan 2 with a 63 × 1.40 NA lens, with binning 2 and a 6% neutral density filter to attenuate the 120W arc mercury epifluorescent source. The motorized filter wheel, external shutters, and the 1,392 × 1,040 pixels 12-bit Photometrics CoolSNAP ES2 camera were controlled by µManager (www.micro-manager. org). Typically, a z-stack of 15 planes 1 µm apart was taken at the 4-cell stage, with exposure times of 50 (DIC), 100 (GFP), and 100 ms (TagRFP-T). For display (Fig. 2a-e; Supplementary  Fig. 1), planes with centriolar signal were retained and max intensity projected in Fiji. Brightness and contrast were adjusted slightly for better visualization, using identical settings within a series.

Indirect immunofluorescence and confocal microscopy
Methanol fixation was performed essentially as described (Gönczy et al. 1999). In brief, gravid hermaphrodites were dissected on polylysine-coated slides, covered with a 12 × 12 mm coverslip, and the slide frozen on a metal block precooled on dry ice. After a few minutes, embryos were freeze-cracked and fixed in −20°C methanol for ∼3 min. Following 2 PBS washes, slides were incubated 45-60 min at room temperature with primary antibodies in PBT (PBS + 0.05% Tween-20). Primary antibodies were 1:50 mouse anti-IFA (Leung et al. 1999) and 1:500 rabbit anti-GFP (a kind gift of Viesturs Simanis). After 2 PBS washes, slides were incubated 45-60 min at room temperature with secondary antibodies in PBT (1:500 goat anti-rabbit-Alexa488 and goat anti-mouse-Alexa568, Thermo Fisher Scientific). Slides were counterstained with 1 µg/ml Hoechst 33258 (Sigma) to reveal DNA and washed twice with PBS. Thereafter, 7 µl of mounting medium (0.189 mol/l n-propyl gallate, 90% glycerol, 10% PBS) was pipetted onto the specimen, which was covered with an 18 × 18 mm coverslip, applying slight pressure to remove excess liquid prior to analysis by wide-field microscopy. Indirect immunofluorescence was imaged on an upright LSM700 Zeiss confocal microscope with a Plan-Apochromat 63 × 1.4 NA lens, collecting 1,024 × 1,024 pixels optical slices 0.35 µm apart, using 405-, 488-, and 555-nm solid state lasers. Relevant planes containing centriolar signals were max intensity projected in Fiji. Brightness and contrast were adjusted slightly for better visualization, using identical settings within a series.

Ethanol fixation
Gravid hermaphrodites were collected from plates with ∼5-ml M9, spun at 1,500 rpm for 2 min in a clinical centrifuge, followed by 2 washes with M9. The supernatant was removed and 1.5-ml 100% ethanol then added. Approximately 1 min thereafter, ethanol was removed and the worm pellet resuspended in 15-µl MVD (50% M9, 50% Vectashield (Vector), 0.7 µg/L Hoechst 33258). After rehydration for 1 min or more, worms were transferred onto a slide and covered with a 20 × 40 mm coverslip, applying slight pressure to remove excess liquid prior to analysis of fixed embryos inside the uterus.

Statistics
The presence of sperm-contributed centrioles in a given blastomere was determined and corresponding contingency tables generated using the Python library pandas (https://pandas. pydata.org/about/citing.html). We then tested whether sister cells (ABa vs ABp, and EMS vs P 2 ) or any of the 4 possible configurations at the 4-cell stage (ABa and P 2 , ABa and EMS, ABp and P 2 , ABp and EMS) were more likely to harbor sperm-contributed centrioles using the chi-square test as implemented in the library SciPy (https://scipy.org/citing-scipy/). An analogous analysis was conducted with embryos that inherited a single spermcontributed centriole. See Supplementary Tables 1 and 2 for details.

Data availability
The authors affirm that all the data necessary for confirming the conclusions of the article are present within the article, figures, and tables. Strains are available upon request. Supplemental material is available at GENETICS online.