EB1 Regulates Tubulin and Actin Cytoskeletal Networks at the Sertoli Cell Blood-Testis Barrier in Male Rats: An In Vitro Study

Abstract

During spermatogenesis, developing germ cells are transported across the seminiferous epithelium. Studies propose that because microtubules (MTs) serve as the tracks for transporting cell organelles, they may also serve a similar function in the transport of developing germ cells. Polarized MTs may provide the tracks along which polarized actin microfilaments, which act as vehicles to transport cargo, such as preleptotene spermatocytes through the blood-testis barrier (BTB) and spermatids across the epithelium. Yet the molecular mechanism(s) underlying these events remain unknown. Using an established in vitro Sertoli cell system to study BTB function, we demonstrated herein that a MT regulatory protein end-binding protein 1 (EB1) regulates the MT- and also the actin-based cytoskeleton of the Sertoli cell BTB in the rat. EB1 serves as a coordinator between the two cytoskeletons by regulating MT polymerization and actin filament bundling to modulate germ cell transport at the Sertoli cell BTB. A knockdown of EB1 by RNA interference was found to perturb the tight junction (TJ)-permeability barrier, as evidenced by mislocalization of junctional proteins critical for barrier function to facilitate spermatocyte transport, which was likely achieved by two coordinated events. First, EB1 knockdown resulted in changes in MT polymerization, thereby perturbing MT organization in Sertoli cells in which polarized MT no longer stretched properly across the cell cytosol to serve as the tracks. Second, EB1 knockdown perturbed actin organization via its effects on the branched actin polymerization-inducing protein called Arp3 (actin-related protein 3), perturbing microfilament bundling capability based on a biochemical assay, thereby causing microfilament truncation and misorganization, disrupting the function of the vehicle. This reduced actin microfilament bundling capability thus perturbed TJ-protein distribution and localization at the BTB, destabilizing the TJ barrier, leading to its remodeling to facilitate spermatocyte transport. In summary, EB1 provides a functional link between tubulin- and actin-based cytoskeletons to confer spermatocyte transport at the BTB.

Spermatogenesis is the process by which diploid spermatogonia differentiate into spermatocytes, which undergo meiosis I/II and develop into haploid spermatids, becoming spermatozoa (1). This process is comprised of a series of tightly regulated hormonal and cellular events that take place within the seminiferous epithelium of the mammalian testis (2–5). The cellular events are largely directed and supported by Sertoli cells, which serve to nourish and structurally support the developing germ cells (3, 6, 7). As they develop, germ cells are progressively transported across the seminiferous epithelium from the basal compartment to the apical compartment. For germ cell transport to occur, cell junctions at the Sertoli-germ cell interface must undergo extensive restructuring (7, 8). Furthermore, spermatids are being transported back and forth across the apical compartment during the epithelial cycle until mature spermatids (ie, spermatozoa) are lined up at the edge of the tubule lumen to prepare for spermiation at late stage VIII of the epithelial cycle (9, 10). Thus, germ cell transport relies almost exclusively on the cytoskeletal networks in Sertoli cells because germ cells per se, in particular spermatids, are metabolically quiescent cells, lacking the locomotive apparatus of other motile cells such as filopodia and lamellipodia (11–13). Therefore, it is not unexpected that Sertoli cells contain extensive actin filament, intermediate filament, and microtubule cytoskeletal networks, all of which serve as scaffolding for the cell and also as structural support for developing germ cells (12–16). The microtubule network is of particular interest because microtubules (MTs) are innately dynamic (12, 13). There are a number of proteins that regulate MT dynamics, ranging from proteins that stabilize and promote polymerization, MT-specific motor proteins, to proteins that sever MTs. It is generally accepted that the dynamic nature of the MT network lends to its critical role in translocation of germ cells, cell shape, and support of developing germ cells. This concept is based on studies in other epithelial cells because there are very few reports in the literature investigating the functional significance of MTs in spermatogenesis, in particular the involvement of MT regulatory proteins in MT dynamics during spermatogenesis.

One of the most widely studied MT regulatory proteins, end-binding protein 1 (EB1), is a regulator of MT dynamics. However, the role of EB1 in the testis remains evasive because there is only one functional study using the testis as a model (17). EB1 belongs to a group of MT regulatory proteins called the plus-end tracking proteins (+TIPs) or end-binding proteins (18–20). Microtubules are polar polymers made up of tubulin subunits, with one end designated the plus end (fast growing end) and the other the minus end (slow growing end). EB1 has been shown to preferentially localize at the plus ends of MTs, usually at cortical sites of a mammalian cell, regulating MT dynamics at that region (19, 21). EB1 promotes MT growth, but the mechanism by which it mediates its effects is controversial; one view is that EB1 promotes catastrophe, whereas another view is that it stabilizes MTs (18, 22). In addition to regulating MT dynamics, EB1 also acts as an adaptor protein that recruits other +TIP proteins to growing MT plus ends (22), illustrating its role in MT dynamics. Although the primary role of EB1 is a regulator of MTs, studies have implicated its function in regulating the actin cytoskeletal network as well (23, 24).

For us to further unfold the biology of spermatid transport during spermatogenesis, a better understanding of microtubule dynamics is necessary. We thus embarked on a study to investigate the role of EB1 in MT organization. If its knockdown by RNA interference (RNAi) in Sertoli cells impedes MT organization, we anticipate that it would also perturb actin microfilament organization and structural integrity. Herein we propose that EB1 plays a critical role in regulating Sertoli cell cytoskeletal dynamics during spermatogenesis at the blood-testis barrier (BTB) through microtubule- and actin-based cell adhesion networks.

Materials and Methods

Animals and antibodies

The use of Sprague Dawley rats (Charles River Laboratories) was approved by the Rockefeller University Institutional Animal Care and Use Committee with protocol number 12–506. Rats were euthanized by CO₂ asphyxiation using slow (20%-30%/min) displacement of chamber air with compressed CO₂. Antibodies used in this report were obtained commercially from different vendors and are listed in Table 1.

Table 1.

Antibodies Used for Different Experiments in This Report

Antibody	Host Species	Vendor	Catalog Number	Working Dilution
				IB/IP	IF/IHC
Actin, polyclonal	Goat	Santa Cruz Biotechnology	sc-1616	1:200 (IB)
Arp3, monoclonal	Mouse	Sigma-Aldrich	A5979	1:3000 (IB)	1:200 (IF)
N-Cadherin, polyclonal	Rabbit	Santa Cruz Biotechnology	sc-7939	1:200 (IB)
CAR, polyclonal	Rabbit	Santa Cruz Biotechnology	sc-15405	1:200 (IB)	1:100 (IF)
β-Catenin, monoclonal	Mouse	Invitrogen	138400	1:300 (IB)	1:100 (IF)
Dia1, polyclonal	Goat	Santa Cruz Biotechnology	sc-10885	1:200 (IB)
EB1, polyclonal	Rabbit	Santa Cruz Biotechnology	sc-15347	1:200 (IB)	1:300 (IF, tissue)
EB1, monoclonal	Mouse	Santa Cruz Biotechnology	sc-374474		1:100 (IHC)
EB1, monoclonal	Mouse	BD Biosciences	610534		1:200 (IF, cells)
Eps8, monoclonal	Mouse	BD Biosciences	610143	1:5000 (IB)	1:100 (IF)
GAPDH, monoclonal	Mouse	Abcam	ab8245	1:1000 (IB)
Katanin p80, polyclonal	Rabbit	Santa Cruz Biotechnology	sc-292216	1:200 (IB)
MARK4, polyclonal	Rabbit	Cell Signaling Technology	4834S	1:500 (IB)
Nectin-3, polyclonal	Goat	Santa Cruz Biotechnology	sc-14806		1:100 (IF)
N-WASP, polyclonal	Rabbit	Santa Cruz Biotechnology	sc-20770	1:25 (IP)
Occludin, polyclonal	Rabbit	Invitrogen	71-1500	1:300 (IB)
Palladin, polyclonal	Rabbit	Protein Tech Group	10853-1-AP	1:1000 (IB)	1:100 (IF)
α-Tubulin, monoclonal	Mouse	Abcam	ab7291	1:1000 (IB)	1:500 (IF/IHC)
β-Tubulin, polyclonal	Rabbit	Abcam	ab6046	1:1000 (IB)	1:500 (IF/IHC)
Vimentin, monoclonal	Mouse	Santa Cruz Biotechnology	sc-6260	1:300 (IB)
ZO-1, polyclonal	Rabbit	Invitrogen	61-7300	1:200 (IB)
ZO-1/FITC conjugated, monoclonal	Mouse	Invitrogen	339111		1:100 (IF)

Abbreviations: FITC, fluorescein isothiocyanate; GAPDH, glyceraldehyde-3-phosphate dehydrogenase; IB, immunoblotting; IP, immunoprecipitation; MARK4, microtubule-associated protein (MAP)/microtubule affinity-regulating kinase 4.

Primary Sertoli cell culture

Sertoli cells were isolated from 20-day-old rat testes, cultured in serum-free F12/DMEM supplemented with growth factors in a humidified atmosphere of 95% air-5% CO₂ (vol/vol) at 35°C as described (25). Sertoli cells were plated on Matrigel (1:7, diluted in F12/DMEM; BD Biosciences)-coated culture dishes, microscopic cover glasses, or bicameral units at different densities, depending on their use for subsequent experiments: 1) 12-well culture dishes at 0.5 × 10⁶ cells/cm² containing 3 mL F12/DMEM per well or six-well culture dishes at 0.4 × 10⁶ cells/cm² containing 5 mL F12/DMEM for lysate preparation or RNA extraction; 2) microscopic (18 mm round) cover glasses at 0.04 × 10⁶ cells/cm² and placed in 12-well dishes containing 2 mL F12/DMEM per well for dual-labeled immunofluorescence analysis; and 3) Millicell bicameral units [Millipore; 12 mm (diameter) inserts, ∼0.6 cm² (effective surface area)] for transepithelial electrical resistance (TER) measurement to assess the Sertoli cell tight junction (TJ) permeability barrier function were placed in 24-well dishes with 0.5-mL F12/DMEM each in the apical and basal compartment. On day 2 after isolation, cells were subjected to a brief hypotonic treatment of 2 minutes, using 10 mM Tris (pH 7.4) to lyse residual germ cells as described (26). Each experiment involving Sertoli cells was repeated at least three times, excluding pilot experiments that assessed optimal experimental conditions.

Assessment of Sertoli cell TJ permeability barrier function

Sertoli cells cultured in vitro are known to establish a functional TJ permeability barrier within approximately 2–3 days, which mimic the Sertoli BTB in vivo with ultrastructures of TJ, basal ectoplasmic specialization (ES; a testis specific adherens junction), gap junction, and desmosome (25, 27, 28). This in vitro system has been widely used by investigators to study Sertoli cell BTB function (29–33). Findings based on the use of this system have been reproduced in vivo (29, 34–36), illustrating its physiological relevance as a BTB study model. To assess Sertoli cell TJ barrier function, each time point had triplicate bicameral units, and each experiment was repeated at least three times using different cell preparations. In short, an approximately 2-second pulse of a 20-μA current was sent across the cell epithelium between two silver-silver chloride electrodes, one of which was placed in the apical compartment of the bicameral unit chamber, and the other in the basal. The resistance that limited the passage of electrical current across the Sertoli cell epithelium because of the presence of the TJ barrier was monitored by using a Millipore MilliCell electrical resistance system. In short, daily (or every 12 h after transfection in selected experiments) reading of TER was obtained in each bicameral unit before fresh F12/DMEM was replenished as described (25).

Transient transfection of small interfering RNA (siRNA) duplexes in cultured Sertoli cells

Sertoli cells were cultured in vitro alone for 3 days to allow the establishment of a functional TJ-permeability barrier. Thereafter, cells were transfected with EB1-specific siRNA vs nontargeting control siRNA duplexes for 24 hours using a TransIT-X2 dynamic delivery system (Mirus Bio LLC) as the transfection medium. The following are the target sequences of the two siRNA duplexes obtained from GE Dharmacon (Fisher Scientific) used to silence EB1: 5′-GGUUAUAAUACACGAGGUG-3′ (ON-TARGETplus Rat Mapre1, J-098686–09-0005) and 5′-CGAAGAAACCUCUCGGCUC-3′ (ON-TARGETplus Rat Mapre1, J-098686–10-0005). A third siRNA duplex (J-098686–11-0005) was found to be ineffective in pilot experiments and was not included in subsequent silencing experiments. Nontargeting siRNA duplex (Silencer Select Negative Control 1 siRNA; Ambion) was used for control experiments.

For immunoblot and immunofluorescence analysis, 50 nM siRNA duplexes was used in treatment vs control groups. However for TER, 100 nM siRNA duplexes was used. Selection of these concentrations of siRNA duplexes for various experiments was based on results of pilot experiments that optimized the experimental conditions to yield detectable phenotypes without cytotoxicity. In brief, on day 3, cells were transfected with siRNA duplexes for a total of 24 hours. After 24 hours, transfection medium was removed, cells were terminated for RNA extraction for RT-PCR (ie, on d 4). However, for cells used for immunoblotting, immunofluorescence, and TER measurement, the transfection medium was removed and replaced with fresh F12/DMEM supplemented with growth factors. Cells for immunoblotting and fluorescence microscopy were incubated for an additional 24 hours before termination for analysis (ie, on d 5). The TER measurement was performed daily. In selected experiments, TER was also quantified every 12 hours after transfection. In some experiments for immunofluorescence analysis, Sertoli cells were cotransfected with 50 nM siRNA duplexes (nontargeting control vs EB1; see above) and 1 nM siGLO red transfection indicator (GE Dharmacon) to assess successful transfection.

MT polymerization assay

A biochemical assay was used to examine changes in microtubule dynamics after in vitro knockdown of EB1 by comparing depolymerized and polymerized MTs between samples. This assay was adapted from earlier studies on tubulin (37–39). Primary Sertoli cells were plated on six-well culture plates at 0.4 × 10⁶ cells/cm² and cultured in F12/DMEM. On day 3, cells were transfected with 50 nM EB1 siRNA duplexes vs nontargeting control siRNA duplexes for 24 hours. Both the treatment and control groups had triplicate wells. Thereafter the transfection medium was removed, replaced with fresh F12/DMEM, and cultured for an additional 24 hours. On day 5, cells were collected in prewarmed lysis buffer (20 mM Tris; 150 mM NaCl; 1% Nonidet P-40; 10% glycerol; 2 mM EGTA; 1 mM MgCl₂, pH 6.9), and cells from both treatment and control groups were processed simultaneously to avoid interexperimental variations. After collection in lysis buffer, the cells were homogenized with a 25-gauge syringe needle and then centrifuged at 130 000 × g at 35°C for 30 minutes to separate free tubulin monomers (supernatant) from microtubules, or polymerized tubulins (pellet). The supernatant was collected, and the pellet was resuspended with the lysis buffer containing 5% sodium dodecyl sulfate to solubilize proteins for SDS-PAGE. The lysates were then processed for immunoblotting to compare the levels of MT polymerization across samples using an anti-β-tubulin antibody (Table 1).

Actin-bundling assay

Actin-bundling activity was assessed using a biochemical assay adapted from a previous study on actin cytoskeletal dynamics (40). Lysates of Sertoli cells transfected with either nontargeting control or EB1 siRNA duplexes were collected with Tris-lysis buffer [20 mM Tris (pH 7.5) containing 20 mM NaCl and 0.5% (vol/vol) Triton X-100] and freshly added protease and phosphatase inhibitors by lysing cells with a 22-gauge needle and syringe, followed by a 28-gauge needle and syringe. The lysates were centrifuged at 20 817 × g for 1 hour at 4°C and subsequently diluted to the same protein concentration. F-actin was prepared using muscle actin (Cytoskeleton, Inc) and diluted to 1 mg/mL with general actin buffer [5 mM Tris-HCl (pH 8.0) containing 0.2 mM CaCl₂ and freshly supplemented 0.2 mM ATP]. The mixture was then incubated at room temperature for 1 hour to obtain 21 μM F-actin stock. Sertoli cell lysate was then added to F-actin stock in a 1:4 ratio of lysate protein to F-actin, followed by 30 minutes of incubation at room temperature to allow for actin-bundling activity. Samples were next centrifuged at 14 000 × g for 5 minutes at 24°C to separate bundled F-actin (pellet) from unbundled F-actin (supernatant). Both the pellet and supernatant were processed via immunoblotting to assess the actin-bundling activity of Sertoli cells after EB1 knockdown.

Dual-labeled immunofluorescence analysis (IF) and immunohistochemistry (IHC)

Dual-labeled IF was performed as described (41) using frozen cross-sections of testes at 7 μm (thickness) obtained at −22°C in a cryostat, or Sertoli cells were cultured on cover glasses at a density of 0.04 × 10⁶ cell/cm². Sections and cells were fixed in either 4% paraformaldehyde in PBS or ice-cold methanol, permeabilized in 0.1% Triton X-100 in PBS, and subsequently blocked in 1% BSA in PBS. Sections were then incubated with specific primary antibodies, listed in Table 1, at appropriate dilution, followed by Alexa Fluor-conjugated secondary antibodies (Alexa Fluor 555 for red fluorescence, Alexa Fluor 488 for green fluorescence; Invitrogen). Negative controls were performed by omitting the primary antibody or substituted with normal IgG and incubating with the respective secondary antibodies. For F-actin staining, sections and cells were incubated with either rhodamine phalloidin (red fluorescence) or fluorescein phalloidin (green fluorescence; Invitrogen). For the staining of nuclei, sections and/or cells were incubated with 50 μg/mL 4′,6-diamidino-2-phenylindole (DAPI) in PBS and then mounted with antifade mounting medium (0.1 M Tris, pH 9.0; n-Propyl gallate; and glycerol).

IHC was performed using Bouin-fixed, paraffin-embedded sections as described (41). In brief, sections were deparaffinized, rehydrated, and then subjected to antigen retrieval in 10 mM citrate buffer (pH 6.0) at 22°C for 10 minutes in a microwave. Sections were blocked with 10% normal serum of the same species as the labeled secondary antibody and then incubated with primary antibody (Table 1) overnight at 4°C. Thereafter, sections were incubated with the respective biotinylated IgG and followed by incubation with streptavidin-horseradish peroxidase (Invitrogen). Color development was performed using a 3-amino-9-ethylcarbazole substrate kit (Invitrogen). Fluorescence and immunohistochemical images were captured using a Nikon Eclipse 90i microscope, and images were acquired using Nikon NIS Elements 3.2 imaging software (Nikon Instruments Inc). If necessary, the images were analyzed using Adobe Photoshop such as image overlay to assess protein colocalization.

Lysate preparation and immunoblotting

Lysates of testes, Sertoli cells, and germ cells were prepared in immunoprecipitation lysis buffer [10 mM Tris (pH 7.4) containing 0.15 M NaCl, 2 mM EGTA, 1% Nonidet P-40 (vol/vol), and 10% glycerol (vol/vol)] supplemented with protease and phosphatase inhibitors freshly added to the lysis buffer at 1 mM 4-(2-aminoethyl)benzene sulfonyl fluoride hydrochloride, 1 mM sodium orthovanadate, 0.05 mM bestatin, 0.05 mM sodium EDTA, 15 μM E64, 1 mM pepstatin, 4 mM sodium tartrate dehydrate, 5 mM NaF, and 3 mM β-glycerophosphate disodium salt. Immunoblotting was performed using a Fujifilm LAS-4000 mini-Luminiscent Image Analyzer and in-house prepared enhanced chemiluminescence kits as described (42).

Coimmunoprecipitation (Co-IP)

To assess the changes in the protein-protein interaction between neuronal Wiskott-Aldrich syndrome protein (N-WASP) and actin-related protein 3 (Arp3) in Sertoli cells after the EB1 knockdown, Co-IP using Sertoli cell lysates (∼300–400 μg protein) and corresponding immunoprecipitation antibody (Table 1) was performed as described (43, 44). In brief, immunocomplexes were precipitated using Protein A/G Plus (Santa Cruz Biotechnology), and protein-protein interaction was then detected by immunoblotting.

RNA extraction and RT-PCR

RNA was extracted from the testes and cells with TRIzol reagent (Invitrogen) for RT-PCR using specific primers against EB1 (Table 2) as described (44).

Table 2.

Primers Used for PCR Experiments in This Report

Gene	Primer Sequence	Orientation	Position	Length, bp	Tm, °C	Cycle Number	GenBank Accession Number
EB1	5′-TCTGCAGTTGAATCTGACAA-3′	Sense	161–180	533	52	30	NM_138509.3
	5′-GCTTCAATACTTTGACCTGC-3′	Antisense	674–693
S-16	5′-TCCGCTGCAGTCCGTTCAAGTCTT-3′	Sense	15–38	385			XM_341815
	5′-GCCAAACTTCTTGGTTTCGCAGCG-3′	Antisense	376–399

Abbreviation: Tm, annealing temperature.

Statistical analysis

Statistical analysis was performed with GB-STAT 7.0 software (Dynamic Microsystems Inc) using a two-way ANOVA. Paired comparison was performed using a Student's t test. All experiments were repeated at least three times using different batches of Sertoli cells.

Results

Cellular distribution of microtubules in the seminiferous epithelium during the epithelial cycle of spermatogenesis

Microtubules are polymers of α-tubulin and β-tubulin heterodimers, which create one of the major cytoskeletal networks of Sertoli cells that support germ cell development including germ cell transport and intracellular organelle trafficking (9, 12, 13). We first used specific antibodies against α-tubulin and β-tubulin (Table 1 and Figure 1, A and B) for IHC (Figure 1C) and IF (Figure 1D) to characterize distribution of α-tubulin in the seminiferous epithelium of adult rat testes during the epithelial cycle of spermatogenesis. The staining pattern of α-tubulin correlates with the structural changes of the seminiferous epithelium during spermatogenesis. Generally, microtubules are distributed longitudinally, spanning across the seminiferous epithelium, from the basal to the apical region, illustrating polarity of microtubules in the seminiferous epithelium, analogous to other mammalian epithelia. Consistent with the fact that α- and β-tubulins are present at an 1:1 equimolar ratio to constitute the tubulin heterodimers, α- and β-tubulins were shown to colocalize in the seminiferous epithelium using frozen sections of adult rat testes (Figure 1E) and also in Sertoli cells cultured in vitro (Figure 1F).

Figure 1.

Cellular distribution and stage-specific expression of α- and β-tubulin in the seminiferous epithelium of adult rat testes. A, B, Specificity of anti-α- and anti-β-tubulin antibodies (Table 1) assessed by immunoblotting using lysate of testes, Sertoli cells (SC), and germ cells (GC). C, Localization of α-tubulin in the seminiferous epithelium shown by IHC analysis using paraffin sections of adult rat testes, illustrating the staining pattern of α-tubulin that reflected structural changes of tubulin-based cytoskeleton that occur throughout the stages of the seminiferous epithelial cycle. Scale bar, 60 μm. D, Expression pattern of microtubules assessed by immunofluorescence microscopy. Fluorescence micrographs present the staining pattern of α-tubulin (green fluorescence), consistent with the pattern shown by IHC shown in panel C. Scale bar, 60 μm. E, Colocalization of α-tubulin (green fluorescence) and β-tubulin (red fluorescence) in which α-/β-tubulin appeared as orange yellow in an early merge micrograph in an early stage VIII tubule. Scale bar, 60 μm. F, Dual-labeled IF analysis to visualize α- and β-tubulin in cultured Sertoli cells, α-tubulin (green fluorescence), β-tubulin (red fluorescence), cell nuclei [blue, 4′,6-diamidino-2-phenylindole (DAPI)]. Scale bar, 50 μm. G, Colocalization of tubulin and F-actin in the seminiferous epithelium of adult rat testes. Dual-labeled immunofluorescence analysis to assess the colocalization of α-tubulin (red fluorescence) and F-actin (green fluorescence) in the seminiferous epithelium of stages V-VIII tubules. Scale bar, 60 μm. H, Dual-labeled IF analysis that illustrated the colocalization of α-tubulin (green fluorescence) and F-actin (red fluorescence) in Sertoli cells. Scale bar, 50 μm.

During spermatogenesis, germ cells must be transported across the BTB and also the seminiferous epithelium because germ cells per se do not possess the necessary apparatus to confer motility, such as lamellipodia and filopodia, thereby relying on the Sertoli cell. Thus, Sertoli cells are highly dynamic cells in which microtubules and the associated motor proteins serve as the track and controllers, whereas actin microfilaments and the specific motor proteins serve as the vehicle and engines to transport germ cells (ie, cargoes) (14, 16, 45–47). As such, it is not surprising that tubulin- and actin-based cytoskeletons are working in concert in the seminiferous epithelium to support the transport of germ cells as well as other intracellular organelles to coordinate the epithelial cycle. Studies in other epithelia also support the concept of cross talk between the microtubule and actin microfilament networks to maintain cellular homeostasis (48, 49). Thus, it is conceivable that cross talk between the two cytoskeletal networks occurs in the testis to coordinate cellular events of spermatogenesis.

Dual-labeled immunofluorescence was used to examine cellular distribution of α-tubulin and F-actin in the seminiferous epithelium. Indeed, α-tubulin partially colocalized with F-actin, conspicuously at the apical ES and the BTB (Figure 1G), which are the actin-rich ultrastructures at the Sertoli-spermatid and Sertoli cell-cell interface, respectively. In stages V and VI of the epithelial cycle, both α-tubulin and F-actin engulfed the heads of developing spermatids; however, by stage VII both cytoskeletal proteins are localized predominantly on the concave (ventral) side of the head of spermatids (Figure 1G). At stage VIII, when the apical ES degenerated to facilitate spermiation, α-tubulin and F-actin expression at the apical ES considerably diminished (Figure 1G). However, the track-like microtubules that extend longitudinally across the seminiferous epithelium remained at stage VII-VIII. Similarly, α-tubulin and F-actin also partially colocalized in cultured Sertoli cells that extend across the cell cytosol (Figure 1H).

EB1 expression in the seminiferous epithelium and in cultured Sertoli cells

Studies have shown that EB1 is a major regulator of microtubule dynamics in multiple epithelia (49–51). In rat testes, EB1 was shown to be expressed by both Sertoli and germ cells in the testis when assessed by immunoblotting (Figure 2, A and B) and RT-PCR (Table 2 and Figure 2C). Studies have shown that EB1 associates with tubulin along the length of microtubules, but it preferentially binds to the plus ends (Figure 2D) (52). Cellular expression of EB1 in the seminiferous epithelium of adult rat testes was assessed by IHC, and it was found at the basal ES/BTB and apical ES (Figure 2E). EB1 is also arranged longitudinally along the epithelium, displaying an expression pattern reminiscent of microtubules in the seminiferous epithelium (Figure 2E vs Figure 1C). EB1 also colocalized with α-tubulin (Figure 2F), consistent with the concept that EB1 is a tubulin binding and tubulin stabilizing protein (Figure 2D). Studies by dual-labeled IF confirmed EB1 was an integrated component of the F-actin-rich anchoring junctions apical ES and basal ES/BTB using markers of the apical ES, namely epidermal growth factor receptor pathway substrate 8 (Eps8), Arp3, and nectin-3, and basal ES such as zonula occludens-1 (ZO-1) and β-catenin (Figure 2G). For instance, EB1 was found predominantly on the concave (ventral) side of the spermatid head at the apical ES and colocalized with both actin regulatory proteins Eps8 and Arp3 but not nectin-3, which was localized mostly to the convex (dorsal) side of the spermatid head at the apical ES (Figure 2G). At the basal ES/BTB, EB1 colocalized with ZO-1, a TJ adaptor protein, and β-catenin, a basal ES adaptor protein (Figure 2G). EB1 expression in cultured Sertoli cells was examined by IF, and similar to other epithelial cells, EB1 was detected at the fast-growing ends, or plus ends, of microtubules appearing as dash-like ultrastructures (Figure 2H), consistent with two earlier reports (17, 53). EB1 also colocalized with MTs in Sertoli cells in vitro, as shown by dual-labeled IF analysis (Figure 2I).

Figure 2.

Cellular and stage-specific localization of EB1 in the seminiferous epithelium of adult rat testes during the epithelial cycle of spermatogenesis. A, Specificity of EB1 antibody (see Table 1) was assessed by immunoblotting using lysate of testes, Sertoli cells (SCs), and germ cells (GCs). Relative level of EB1 in lysate from adult rat testes (T), SCs isolated from 20-day-old rat testes, and GCs isolated from adult rat testes with actin serving as an internal loading control. B, Composite immunoblot data were shown in this histogram. Each bar is a mean ± SD of three independent experiments. C, EB1 steady-state mRNA level in T, SCs, and GCs was also assessed by RT-PCR with S-16 serving as an internal control. The authenticity of mRNA transcripts detected was confirmed by nucleotide sequencing. D, An illustration of EB1 distribution along a microtubule that is composed of polymerized α-tubulin (blue) and β-tubulin (red) heterodimers stabilized by EB1. EB1 preferentially binds to the plus (+) end of a microtubule but is also found along the length of the microtubule and at the minus (−) end. E, Localization of EB1 in the seminiferous epithelium throughout stages I-XIV of the seminiferous epithelial cycle as illustrated by IHC analysis using paraffin sections of adult rat testes. Scale bar, 60 μm. Ctrl, Control. F, Select stages, V and VII, that depict the colocalization of EB1 (red) and α-tubulin (green) by dual-label IF analysis. Scale bar, 60 μm. DAPI, 4′,6-diamidino-2-phenylindole. G, Colocalization of EB1 (red) with apical ES (green) proteins Eps8, Arp3, and nectin-3 and also of EB1 (red) with basal ES/BTB (green) proteins ZO-1 and β-catenin. Scale bar, 40 μm for apical ES proteins; 50 μm for basal ES proteins. H, Cellular distribution of EB1 in cultured Sertoli cells. The boxed area was magnified and shown on the right panel in which EB1 was detected at the fast growing ends, or plus ends, of microtubules that appeared as dash-like ultrastructures. Scale bar, 20 μm; 8 μm in magnified image. I, Dual-labeled IF analysis of colocalization of EB1 (green) and β-tubulin (red) in Sertoli cells. Scale bar, 60 μm.

Knockdown of EB1 by RNAi perturbs Sertoli cell TJ permeability barrier in vitro via changes in adhesion protein distribution at the cell-cell interface

An in vitro system composed of purified Sertoli cells with a functional TJ barrier that mimicked the Sertoli cell BTB in vivo as earlier described (7, 28, 54) was used to study the functional role of EB1 in the testis. EB1 was silenced by approximately 80% using siRNA duplexes specific for EB1 (Figure 3, A and B). Overall, the effect of knockdown of EB1 on steady-state levels of selected proteins crucial for the maintenance of cytoskeletal and cell adhesion at the BTB were unaffected as demonstrated by immunoblot analysis (Figure 3A). Results show that knockdown of EB1 perturbed the Sertoli cell TJ permeability barrier (Figure 3C). This TJ barrier disruption is mediated via changes in the localization of adhesion proteins at the TJ, such as coxsackievirus and adenovirus receptor (CAR) and ZO-1, and at the basal ES, such as N-cadherin but not β-catenin after the knockdown of EB1 by approximately 80% (Figure 3D). For instance, these TJ and basal ES proteins were redistributed and moved to the cell cytosol instead of being localized to the Sertoli cell-cell interface, thereby destabilizing BTB integrity.

Figure 3.

Knockdown of EB1 by RNAi in Sertoli cell epithelium with a functional BTB in vitro perturbs the TJ permeability barrier mediated by the mislocalization and/or redistribution of BTB-associated proteins. A, Sertoli cells with an established TJ permeability barrier were transfected with EB1 siRNA duplexes to knock down EB1 vs nontargeting control duplexes on day 3 for 24 hours. Immunoblotting analysis using lysates obtained from cells terminated on day 5 was performed with specific antibodies (Table 1) against selected target proteins at the BTB. β-Actin and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) served as protein-loading controls. B, A knockdown of EB1 by approximately 80% was noted in which the control was arbitrarily set to 1 against which statistical comparison was performed. Each bar is a mean ± SD of three independent experiments. Except for EB1, the steady-state levels of all other proteins were not affected following EB1 knockdown, illustrating no apparent off-target effects using the EB1 siRNA duplexes for silencing. **, P < .01 by Student's t test. Ctrl, control, C, In parallel experiments, Sertoli cells cultured on bicameral units with an established TJ permeability barrier were transfected with EB1 siRNA duplexes vs nontargeting control siRNA to monitor changes in the barrier function after the EB1 knockdown. **, P < .01 by Student's paired t test. D, The expression of select TJ barrier proteins after silencing of EB1 was visualized via immunofluorescence microscopy. The siGLO transfection indicator (red; Dharmacon) indicates a successful transfection. Scale bar, 25 μm, which applies to other micrographs.

EB1 knockdown affects microtubule organization in Sertoli cells

EB1 knockdown in Sertoli cells was found to affect microtubule organization in these cells (Figure 4A). In controls in which cells were transfected with nontargeting siRNA duplexes, most microtubules extended throughout the cytoplasm, radiating from the cell nucleus toward the cell periphery (Figure 4A). After EB1 knockdown, Sertoli cell microtubules, including both β-tubulin (Figure 4A) and α-tubulin (Figure 4B) exhibited a denser tubulin localization surrounding the nucleus (Figure 4, A and B). To analyze these findings semiquantitatively, microtubule extension from the cell nucleus toward the cell periphery was quantified by taking the average of the shortest (see white bracket in Figure 4, A and B) and longest (see yellow bracket in Figure 4, A and B) microtubule extension in cells from both groups (Figure 4C). After EB1 knockdown, there is a significant reduction in microtubule extension, approximately 40%, in Sertoli cells (Figure 4C), supporting the notion that microtubules were more concentrated around the nucleus than throughout the cytoplasm because reduced EB1 failed to stabilize microtubules to maintain their robust polymerization, generating shorter microtubules throughout the cell. To further confirm these observations, a biochemical assay that estimated the ability of microtubule/tubulin polymerization was performed after EB1 knockdown vs control cells (Figure 4D). Sertoli cell lysate from both groups were subjected to ultracentrifugation to separate polymerized microtubules (pellet) and free tubulin monomers (supernatant). It was revealed that EB1 knockdown perturbed the ability of microtubule polymerization as evidenced by a considerable decline in the β-tubulin protein level in the pellet that assessed polymerized microtubules, concomitant with a mild increase in free tubulin monomers in supernatant in the EB1 knockdown cells (Figure 4D).

Figure 4.

Changes in the microtubule-based cytoskeleton in cultured Sertoli cells after EB1 knockdown by RNAi. A and B, IF analysis to assess the changes in the localization and expression of microtubules (red fluorescence or gray scale for β-tubulin in panel A, green fluorescence or gray scale for α-tubulin in panel B) after EB1 KD (EB1, green fluorescence). For panel A, scale bar, 30 μm, top panel; 25 μm, middle and bottom panels. For panel B, scale bar, 100 μm, top panel; 25 μm, middle and bottom panels. Ctrl, control; DAPI, 4′,6-diamidino-2-phenylindole. C, Histograms to provide semiquantitative analysis depicting changes in tubulin-based cytoskeleton after the knockdown of EB1 vs controls at the Sertoli cell BTB. Each bar is a mean ± SD of the distance of the longest (see yellow bracket) and the shortest (see white bracket) microtubules radiating from the Sertoli cell nucleus of 150 cells from three independent experiments (50 randomly selected Sertoli cells in each experiment), supporting the conclusion that after the EB1 knockdown, the tubulin-based cytoskeleton retracted from the Sertoli cell cytosol by moving closer to the cell nucleus as the result of a reduced MT polymerization. D, A tubulin polymerization assay was performed to detect any changes in polymerization dynamics of the MT network within Sertoli cells after the silencing of EB1. Pellet contains tubulin polymers, whereas the supernatant contains free tubulin monomers. A loss in tubulin polymerization in EB1 knockdown Sertoli cells was noted vs controls.

Knockdown of EB1, a microtubule regulatory protein, perturbs actin organization in Sertoli cells

Current studies support the concept of cross talk between actin- and microtubule-based cytoskeleton mediated by EB1 because it is localized predominantly at the tips (plus ends) of microtubules, placing the protein at an ideal location to interact with proteins at the cell cortex and cell-cell interface (49). After the knockdown of EB1 by approximately 80% (Figure 5, A, and B), the spatiotemporal localization of a branched actin polymerization-inducing protein Arp3 was considerably altered but not actin microfilament barbed-end capping and bundling protein Eps8 or actin cross-linking and bundling protein palladin (Figure 5, A and C). It is noted that when activated by N-WASP, the Arp2/3 complex induces branched actin polymerization, effectively conferring actin-based cytoskeleton plasticity so that actin microfilaments can be reorganized from their bundled to branched configuration. In control cells, Arp3 is abundantly localized to the cell-cell interface, with some Arp3 in the cell cytosol near the nucleus. However, after EB1 knockdown, the localization of Arp3 at the cell-cell interface considerably diminished but increased in the cell cytosol around the nuclear region (Figure 5A), and fluorescence intensity of Arp3 at the Sertoli cell-cell interface and near the nuclear region was quantified and compared between EB1 knockdown vs control cells (Figure 5C). To further examine whether such changes in the cellular distribution in Arp3 would indeed affect branched actin polymerization, we assessed whether there was an increase in an N-WASP-Arp3 association after the EB1 knockdown. A study by Co-IP indeed confirmed an increase in N-WASP and Arp3 association after the EB1 knockdown (Figure 5D). Findings shown in Figure 5D seemingly suggested that such an increase in the Arp2/3 activation by N-WASP would impede actin microfilament organization in Sertoli cells.

Figure 5.

The EB1 knockdown-induced Sertoli cell TJ permeability barrier disruption is mediated by changes in the spatiotemporal expression and/or distribution of actin regulatory proteins Arp3, Eps8, and palladin at the Sertoli cell BTB. A, IF analysis to confirm the knockdown of EB1 (green) in Sertoli cells by approximately 80% after the transfection of EB1-specific siRNA duplexes (EB1 RNAi) vs nontargeting control duplexes (Ctrl RNAi) and to examine the distribution of actin regulatory proteins before and after EB1 knockdown. Scale bar, 25 μm, which applies to other micrographs. Ctrl, control; DAPI, 4′,6-diamidino-2-phenylindole. B, Image analysis of EB1 fluorescence signals in Sertoli cells transfected with EB1 specific siRNA duplexes vs. nontargeting control siRNA duplexes (see A) to confirm the knockdown of EB1 by ∼80%. C, Changes in the relative distribution of Arp3 at the Sertoli cell-cell interface vs the nuclear region were compared in this histogram by quantifying the fluorescence signals at both sites in 120 randomly selected cells from three independent experiments with approximately 40 cells per experiment. The fluorescence signal of Arp3 in cells from the Ctrl RNAi group at both the cell-cell interface and nuclear region was arbitrarily set at 1, against which the statistical comparison was performed. Each bar is a mean ± SD of three independent experiments. **, P < .01. D, Co-IP analysis using anti-N-WASP IgG (see Table 1) was used to assess changes in the interaction between N-WASP and Arp3 after EB1 knockdown. The histogram in the lower panel is a summary of the data shown in the upper panel. Each bar is a mean ± SD of three experiments. *, P < .05.

This postulate was indeed confirmed when F-actin organization in the EB1 knockdown cells was assessed vs control cells because actin microfilaments in EB1 knockdown cells were found to become truncated, no longer properly organized across the cell cytosol as highly polarized undisrupted microfilaments shown in control Sertoli cells (Figure 6A). These findings thus support the findings shown in Figure 3 in which the knockdown of EB1 perturbed the Sertoli cell TJ-barrier due to a mislocalization of adhesion proteins at the Sertoli cell BTB because both adhesion proteins CAR and N-cadherin use actin microfilaments for attachment. To further confirm these observations, a biochemical assay was used to compare the actin-bundling activity of the EB1 knockdown cells vs control cells (Figure 6B). The results showed that the knockdown of EB1 significantly reduced the ability of the Sertoli cells to form actin bundles (Figure 6B).

Figure 6.

Changes in the organization of actin microfilaments at the Sertoli cell BTB after knockdown of EB1. A, Immunofluorescence analysis was performed after EB1 (green fluorescence) knockdown to visualize changes in the cellular distribution of actin microfilaments (gray scale or green fluorescence) in Sertoli cells. Actin microfilaments in Sertoli cells after EB1 knockdown were truncated and disorganized, no longer stretched out to encompass the entire Sertoli cell cytosol, and part of the Sertoli cell nucleus was not covered with actin microfilaments, distinctively different from cells treated with nontargeting control siRNA duplexes. Scale bar, 25 μm, which applies to other micrographs. Ctrl, control; DAPI, 4′,6-diamidino-2-phenylindole. B, Actin-bundling assay was performed to support findings shown in panel A regarding changes in F-actin organization in Sertoli cells after the EB1 knockdown. Knockdown of EB1 was shown to reduce actin-bundling capability, and these changes were significant (lower panel). Actin-bundling activity in Ctrl RNAi Sertoli cells was arbitrarily set at 1. Each bar in the histogram is a mean ± SD of three independent experiments. *, P < .05.

Discussion

EB1, a +TIP, is a MT-stabilizing protein located mostly at the growing plus ends of MTs in the testis as shown herein, consistent with findings in other epithelia, which is important for cell division, cell shape/form, cell movement, cell survival, and intracellular endocytic vesicle/organelle trafficking (19, 51). However, studies have shown that EB1 is also localized along MTs because EB1 was found to colocalize with α- or β-tubulin in the seminiferous epithelium in vivo and also in the cytosol of Sertoli cells cultured in vivo, illustrating that it is also being used to stabilize MTs, analogous to its role in MTs in other epithelia and mammalian cells (19, 51). Yet the mechanism(s) by which EB1 exerts its function to modulate MT dynamics remains elusive. In motile cells, such as osteoclasts, EB1 was recently shown to work in concert with cortactin and Src to confer the dynamic nature of podosomes (55), which are short-life span, F-actin-rich dynamic structures (∼2–4 min) supported by MTs in motile cells such as oesteoclasts, macrophages and dendritic cells that facilitate locomotion of mammalian cells on extracellular matrix, such as bone surface and basal lamina (56–58). It was shown that cortactin is an Src-dependent binding partner of EB1, and EB1 depletion by RNAi led to a loss of a podosome belt in osteoclasts, with a concomitant increase in cortactin phosphorylation that affected cortactin interaction with EB1, perturbing bone resorption (55).

It is of interest to note that Sertoli cells are motile cells when cultured in vitro, capable of traversing the membranes of bicameral units, moving from the apical to the basal compartment, analogous to metastatic tumor cells (59, 60). However, Sertoli cells are highly polarized cells in the seminiferous epithelium in which the cell nuclei are localized restrictively to the basal compartment near the basement membrane of tunica propria (7, 61). Even though Sertoli cells are cyclic in nature in reference to their secretory and metabolic activity in response to different stages of the epithelial cycle (3, 6), they remain relatively static in the epithelium because each Sertoli cell has to support approximately 30–50 developing germ cells (62). On the other hand, developing germ cells, most notably postmeiotic spermatids, are metabolically quiescent cells, lacking the ultrastructures of podosomes, filopodia, and lamellipodia found in other motile cells. Thus, it is unlikely that EB1 exerts its effects in MTs to affect Sertoli cell movement analogous to osteoclasts, macrophages, and dendritic cells. However, it is likely that EB1 in MTs exerts its function to track the growing plus end of MTs to allow transport of junction components (eg, integral membrane proteins, adaptors), organelles (eg, endosomes, phagosomes) via endocytic vesicle-mediated trafficking pathways necessary to support spermatogenesis as well as germ cell transport during the epithelial cycle.

This notion is supported by the current findings that a knockdown of EB1 via RNAi by approximately 80% in Sertoli cells with an established TJ barrier would impede adhesion protein localization at the BTB, perturbing the barrier function. For instance, the localization of both the TJ proteins CAR and ZO-1 as well as basal ES protein N-cadherin were perturbed, thereby perturbing the Sertoli TJ permeability barrier function. This is likely mediated by endocytic vesicle-mediated protein trafficking because EB1 is known to work in concert with c-Src (55), and a recent report has shown that c-Src is a crucial regulator of protein endocytosis, transcytosis, and recycling at the Sertoli cell BTB (63). Consistent with the role of EB1 in MT dynamics (19, 51), the knockdown of EB1 was found to reduce MT polymerization based on a biochemical assay, thereby perturbing the growth of MTs in Sertoli cells, reducing the ability of MTs to stretch along the Sertoli cell, ie, reducing the growth of MTs because of the lack of this crucial MT stabilizing protein.

Because EB1 is a +TIP, it thus places this protein at a location to interact with actin microfilaments at the cell cortex and cell-cell interface as noted in studies of other epithelia (49). Herein we demonstrated in the testis that a knockdown of EB1 in Sertoli cells by approximately 80% caused significant changes in the organization of F-actin at the BTB in which actin microfilaments were found to become truncated and disorganized, no longer orderly spanning the Sertoli cell cytosol to support cell adhesion complexes (eg, CAR-ZO-1) at the Sertoli cell-cell interface. More importantly, it was noted that these changes were mediated by an alteration of organization of the actin-regulatory proteins, such as Arp3 (64) at the Sertoli cell-cell interface. Although the localization of actin barbed-end capping and bundling protein Eps8 (65) and actin cross-linking and bundling protein palladin (44) was unaffected, a mislocalized Arp3 and a reduced Arp3/N-WASP interaction thus impeded proper organization of actin microfilaments to confer adhesion proteins at the cell-cell interface to maintain the barrier function. In short, EB1 also stabilizes actin microfilaments that lie near MTs in Sertoli cells to promote BTB function via its stabilizing function with actin regulatory proteins.

In summary, we provide unequivocal evidence that EB1 is a crucial regulator of MTs in Sertoli cells by stabilizing the growth of MTs and MT maintenance. However, EB1 also confers actin microfilament organization and integrity via its effects on the Arp2/3 complex at the Sertoli cell BTB. It remains to be determined whether EB1 mediates its effects via its interacting partners c-Src, cortactin, and others.