UNCORRECTED PROOF
Experimental Cell Research xxx (2016) xxx-xxx
Contents lists available at ScienceDirect
Experimental Cell Research
journal homepage: www.elsevier.com
Gamma-tubulin coordinates nuclear envelope assembly around chromatin
Catalina Ana Rosselló,
Lisa Lindström,
Johan Glindre,
Greta Eklund,
Maria Alvarado-Kristensson
Division of Molecular Pathology, Department of Translational Medicine, Lund University, Skåne University Hospital, Malmö, 20502. Sweden
Division of Molecular Pathology Department of Translational Medicine Lund University Skåne University Hospital Malmö 20502 Sweden
Keywords: Bi­o­log­i­cal sci­ences; Cell bi­ol­ogy
Abstract
The cy­toso­lic role of γ-tubu­lin as a mi­cro­tubule or­ga­nizer has been stud­ied thor­oughly, but its nu­clear func­tion is poorly un­der­stood. Here, we show that γ-tubu­lin is lo­cated through­out the chro­matin of de­mem­branated Xeno­pus lae­vis sperm and, as the nu­cleus is formed, γ-tubu­lin re­cruits lamin B3 and nu­clear mem­branes. Im­mun­ode­ple­tion of γ-tubu­lin im­pairs X. lae­vis as­sem­bly of both the lam­ina and the nu­clear mem­brane. Dur­ing nu­clear for­ma­tion in mam­malian cell lines, γ-tubu­lin es­tab­lishes a cel­lu­lar pro­tein bound­ary around chro­matin that co­or­di­nates nu­clear as­sem­bly of the daugh­ter nu­clei. Fur­ther­more, ex­pres­sion of a γ-tubu­lin mu­tant that lacks the DNA-bind­ing do­main forms chro­matin-empty nu­clear like struc­tures and demon­strate that a con­stant in­ter­play be­tween the chro­matin-as­so­ci­ated and the cy­toso­lic pools of γ-tubu­lin is re­quired and, when the bal­ance be­tween pools is im­paired, aber­rant nu­clei are formed. We there­fore pro­pose that the nu­clear pro­tein mesh­work formed by γ-tubu­lin around chro­matin co­or­di­nates nu­clear for­ma­tion in eu­kary­otic cells.
1. Introduction
The chain of events that leads to the for­ma­tion of two iden­ti­cal daugh­ter cells in cell di­vi­sion is highly reg­u­lated, with syn­chro­niza­tion of cy­toso­lic and nu­clear events. Preser­va­tion of cel­lu­lar home­osta­sis and ge­nomic in­tegrity re­quires the co­or­di­nated mod­i­fi­ca­tion of cell com­po­nents dur­ing cell di­vi­sion. One such com­po­nent is the nu­clear en­ve­lope, which sep­a­rates the chro­mo­somes from cy­to­plas­mic struc­tures such as the cen­tro­somes that nu­cle­ate spin­dle mi­cro­tubules. At the on­set of mi­to­sis the nu­clear en­ve­lope dis­as­sem­bles to al­low the mi­totic spin­dle to seg­re­gate the con­densed chro­mo­somes be­tween daugh­ter cells (Collas, 1999; Peter et al., 1990) and re­assem­bles at anaphase/​telophase (Thompson et al., 1997). The nu­clear lam­ina is a struc­tural com­po­nent of the in­ner sur­face of the nu­clear en­ve­lope that pro­vides sup­port for chro­matin do­mains and nu­clear en­ve­lope pro­teins (Dechat et al., 2010) and in hu­mans con­tains three ma­jor struc­turally re­lated pro­teins named lamin A, B and C (Dechat et al., 2010). At the on­set of mi­to­sis, dis­perse lamin B re­mains as­so­ci­ated with both the nu­clear en­ve­lope mem­brane frag­ments and the mi­totic spin­dle in the cy­to­plasm (Tsai et al., 2006). How­ever, the mech­a­nism by which nu­clear as­sem­bly en­sures the for­ma­tion of two diploid daugh­ter nu­clei at the end of mi­to­sis is poorly un­der­stood.
γ-Tubu­lin is a cy­toso­lic pro­tein that reg­u­lates α- and β-tubu­lin nu­cle­ation to a grow­ing mi­cro­tubule (Kollman et al., 2011). We and oth­ers have shown that nu­clear γ-tubu­lin as­so­ci­ates with Rad51, C53, GCP2, GCP3 and E2F1 (Draberova et al., 2015; Ehlen et al., 2012; Hoog et al., 2011; Horejsi et al., 2012; Lesca et al., 2005) and in­hi­bi
Corresponding author at: Maria Alvarado-Kristensson, Molecular Pathology, Lund University, SUS-MAS, 59 Jan Waldenströms Street, Floor 2, SE-205 02 Malmö, Sweden.
Email address: maria.​alvarado-kristensson@​med.​lu.​se (M. Alvarado-Kristensson)
­tion of the nu­clear ac­tiv­ity of γ-tubu­lin does not in­ter­fere with mi­cro­tubule dy­nam­ics (Lindstrom et al., 2015). Dur­ing the cell cy­cle, γ-tubu­lin has var­i­ous func­tions, reg­u­lat­ing cen­tro­so­mal du­pli­ca­tion and mi­totic spin­dle for­ma­tion and mod­er­at­ing tran­scrip­tional ac­tiv­i­ties of E2Fs dur­ing S-phase (Hoog et al., 2011).
Al­though pre­vi­ous work has high­lighted the role of mi­cro­tubules and of the γ-tubu­lin com­plex pro­tein 3-in­ter­act­ing pro­teins in shap­ing the nu­clear en­ve­lope (Batzenschlager et al., 2013; Xue et al., 2013), it is not known whether nu­clear for­ma­tion and nu­clear γ-tubu­lin are func­tion­ally linked. In the pre­sent study, we show, us­ing both Xeno­pus lae­vis ex­tracts (Lohka and Masui, 1983) and mam­malian cell lines, that γ-tubu­lin forms a cel­lu­lar mesh­work around chro­matin and has two main roles dur­ing nu­clear as­sem­bly. First, γ-tubu­lin builds up a nu­clear pro­tein bound­ary that con­nects the cy­to­plasm and the nu­clear com­part­ment to­gether, and, sec­ond, γ-tubu­lin co­or­di­nates nu­clear for­ma­tion.
2. Results
2.1. Chromatin-associated γ-tubulin is necessary for nuclear assembly
In cell lines, TUBU­LIN-shRNA ex­pres­sion only par­tially re­duces the γ-tubu­lin pro­tein lev­els (Eklund et al., 2014; Hoog et al., 2011), mak­ing de­ple­tion ex­per­i­ments dif­fi­cult to in­ter­pret. γ-Tubu­lin is highly con­served among species; at the pro­tein level hu­man and X. lae­vis γ-tubu­lin-1 are 98% ho­mol­o­gous (Fig. 1A). For this rea­son, X. lae­vis egg ex­tracts have ex­ten­sively been used to study the role of γ-tubu­lin in mi­cro­tubule nu­cle­ation (Felix et al., 1994).
Analy­sis of de­mem­branated X. lae­vis sperm showed that as hi­s­tone 3, γ-tubu­lin lo­cal­ized through­out the DNA (Fig. 1B), whereas α-
UNCORRECTED PROOF
Fig. 1. De­mem­branated sperm con­tains γ-tubu­lin. (A) Se­quence align­ment of the vari­able re­gion of hu­man γ-tubu­lin 1 and γ-tubu­lin 2 and Xeno­pus lae­vis γ-tubu­lin, show­ing residues 334451 of the C ter­mi­nal re­gion. Bold let­ters in­di­cate dif­fer­ences. (B, C) The lo­cal­iza­tion of γ-tubu­lin, hi­s­tone 3, γ-tubu­lin ring com­plex (Xgrip109) and cen­tro­some (cen­trin and αTub) were im­muno­flu­o­res­cence stained in de­mem­branated sperm (n = 5) with an anti-γ-tubu­lin an­ti­body that was pro­duced ei­ther in rab­bit (γ-TubR; T3320) or mouse (γ-TubM; ab27074). (Da) To­tal lysates of egg ex­tract and de­mem­branated sperm were an­a­lyzed by west­ern blot­ting with the in­di­cated an­ti­bod­ies (n = 3; T3320). An actin and α-tubu­lin load­ing con­trols are shown. (Db) The in­di­cated amount of egg ex­tract (egg) and de­mem­branated sperm (sperm) was an­a­lyzed by west­ern blot­ting with the in­di­cated an­ti­bod­ies (n = 6; T3320). An α-tubu­lin load­ing con­trol is shown. (E) The anti-γ-tubu­lin an­ti­body pro­duced in rab­bit or mouse (T3320 and ab27074) were prein­cu­batd for 2 h with Ni2+ affin­ity resin (con­trol) or with Ni2+ affin­ity resin as­so­ci­ated with His6-tagged γ-tubu­lin (ab­sorb. His6-γ-Tub) be­fore im­muno­flu­o­res­cence stain­ing the sperm. In (B, C, E) γ-tubu­lin is shown as green and hi­s­tone 3 as red and nu­clei as blue (DAPI). In (C) Xgrip109 and γTub are shown as green and red, re­spec­tively. Scale bars, 10 μm.
alt-text: Fig. 1
tubu­lin and γ-tubu­lin ring com­plex pro­tein, Xgrip109 (Martin et al., 1998), were found in one end of the sperm (Fig. 1C). Si­mul­ta­ne­ous stain­ing of α-tubu­lin and cen­trin showed the cen­tro­so­mal lo­cal­iza­tion of αβ-tubu­lin dimers (Fig. 1C) and ex­plained the pres­ence of αβ-tubu­lin in the sperm (Fig. 1D). No­tably, the ob­served con­sti­tu­tive lo­ca­tion of en­doge­nous γ-tubu­lin to the chro­matin of sperm was re­duced upon prein­cu­ba­tion of the anti-γ-tubu­lin an­ti­body (T3320 or ab27074) with Ni2+ affin­ity resin as­so­ci­ated His6-tagged γ-tubu­lin be­fore im­muno­flu­o­res­cence stain­ing the sperm (Fig. 1E), demon­strat­ing the speci­ficity of the im­muno­flu­o­res­cence stain­ing. West­ern blot analy­sis of γ-tubu­lin con­firmed that both egg ex­tracts and sperm con­tained γ-tubu­lin, al­though, the γ-tubu­lin to α-tubu­lin ex­pres­sion ra­tio was ap­prox­i­mately nine to sev­en­teen fold higher in the sperm (13.0 ± 8.2; n = 7) (Fig. 1D).
To ex­clude an in­volve­ment of αβ-tubu­lin in the ini­tial events lead­ing to nu­clear for­ma­tion, we pre­pared the sperm in the pres­ence of col­cemid and re­moved αβ-tubu­lin de­bris by a glyc­erol cush­ion (Fig. 2A) (Felix et al., 1994). This treat­ment re­duced the amount of αβ-tubu­lin and cen­trin as­so­ci­ated to the sperm (Fig. 2A). Ad­di­tion of egg ex­tracts to the col­cemid pre­treated sperm trig­gered nu­clear for­ma­tion (Fig. 2B) (Lohka and Masui, 1983). Fur­ther­more, 100% of the formed nu­clei ex­cluded TRITC-la­beled 155-kDa dex­tran (99.7 ± 0.6; n = 3) and as­sem­bled nu­clear pore com­plexes, con­firm­ing the in­tegrity of the nu­clei (Fig. 2C).
Dur­ing nu­clear as­sem­bly, the sperm un­dergo four dis­tinct mor­pho­log­i­cal stages be­fore form­ing a nu­clear γ-tubu­lin bound­ary (Fig. 2B). In stage 1, the sperm is con­densed and in stage 2, de­con­densed. An al­most formed nu­cleus is ob­served in stage 3 and is larger dur­ing stage 4. Chro­matin-as­so­ci­ated γ-tubu­lin was lo­cal­ized through­out the
sperm chro­matin in stage 1 to 3 and en­riched at the nu­clear en­ve­lope (NE) in stage 4 (Fig. 2B), but was not in com­plex with Xgrip109 (Fig. 2D).
To find func­tional links be­tween αβ-tubu­lin and chro­matin-as­so­ci­ated γ-tubu­lin, we stud­ied the lo­ca­tion of α-tubu­lin dur­ing nu­clear for­ma­tion and found that γ- and α-tubu­lin co­in­cided in cen­tro­somes (Fig. 2E) (Xue et al., 2013). To in­ves­ti­gate the role of αβ-tubu­lin, we per­formed nu­clear as­sem­bly re­ac­tions in the pres­ence of var­i­ous con­cen­tra­tions of col­cemid and as­sayed its ef­fect on the for­ma­tion of en­doge­nous mi­cro­tubules. Sup­ple­men­ta­tion with 100 ng/​ml col­cemid de­poly­mer­ized mi­cro­tubules (Fig. 2F), but nei­ther im­paired nu­clear for­ma­tion, nor af­fected the lo­ca­tion of γ-tubu­lin, sug­gest­ing that γ-tubu­lin has mi­cro­tubule-in­de­pen­dent func­tions dur­ing nu­clear as­sem­bly (Fig. 2E, F).
Im­mun­ode­ple­tion of γ-tubu­lin from egg ex­tracts re­duced the amount of γ-tubu­lin by 54 ± 6% (Fig. 3A; n = 6) and caused a trend to­wards de­crease nu­clear for­ma­tion (Fig. 3B). To fur­ther im­prove the de­gree of γ-tubu­lin im­mun­ode­ple­tion in the sperm (Fig. 1B-E; stage 1), we re­moved chro­matin-bound pro­teins with pro­teinase K (Ksperm; Fig. 3C) and found that only de­ple­tion of γ-tubu­lin in both sperm and egg ex­tracts sig­nif­i­cantly im­paired nu­clear for­ma­tion (Fig. 3A-D). Fur­ther­more, sup­ple­men­ta­tion of the egg ex­tracts with re­com­bi­nant γ-tubu­lin par­tially re-es­tab­lished nu­clear for­ma­tion, but it was not as ef­fi­cient as when added to the sperm (Fig. 4A). In­deed, sup­ple­men­ta­tion of Ksperm with re­com­bi­nant γ-tubu­lin-1 or the tran­scrip­tion fac­tor E2F1 mu­tant, E2F1Δ194426 (Hoog et al., 2011) proved that only γ-tubu­lin re-es­tab­lished nu­clear for­ma­tion (Fig. 4A). These data im­ply that γ-tubu­lin is nec­es­sary for nu­clear as­sem­bly.
UNCORRECTED PROOF
Fig. 2. γ-Tubu­lin and α-tubu­lin are dif­fer­ently dis­trib­uted. (A) γ-Tubu­lin (T5192), α-tubu­lin and cen­trin were im­muno­flu­o­res­cence stained in de­mem­branated sperm that were iso­lated in the pres­ence of col­cemid and pel­leted onto a cush­ion con­tain­ing SuNaSp and glyc­erol be­fore the im­munos­tain­ing (n = 5). (B) To study mor­pho­log­i­cal changes, de­mem­branated sperm were in­cu­bated with egg ex­tracts dur­ing 0 min (stage 1, con­densed sperm) or 60 min be­fore fix­a­tion (stage 2, de­con­densed; stage 3, small al­most-formed nu­cleus; and stage 4, nu­cleus with a γ-tubu­lin bound­ary). Lo­cal­iza­tion of γ-tubu­lin was ex­am­ined by im­muno­flu­o­res­cence stain­ing with a rab­bit and a mouse pro­duced anti-γ-tubu­lin an­ti­body, as in­di­cated. Graph shows the mean per­cent­age of formed nu­clei (open bar; stage 3 and 4; ± s.d., n = 3; * p < 0.05). (C) To test the in­tegrity of formed nu­clei, nu­clear as­sem­bly re­ac­tions were in­cu­bated with TRITC-la­beled dex­tran. Al­ter­na­tively, nu­clear pore com­plexes (NPC) were im­muno­flu­o­res­cence stained (n = 3). (D-F) De­mem­branated sperm pre­pared as in Fig. 1B (D, E) or as in A (F) were in­cu­bated in the pres­ence of egg ex­tracts for 60 min be­fore fix­a­tion in the ab­sence (D, E) or pres­ence (F) of col­cemid. The lo­cal­iza­tion of Xgrip109, mi­cro­tubules (αTub), cen­tro­somes (γTub and αTub) and γ-tubu­lin were ex­am­ined by im­muno­flu­o­res­cence stain­ing (n = 3). (F) From left to right, first graph shows the mean per­cent­age of formed nu­clei in the pres­ence of col­cemid (black bar) rel­a­tive to a non-treated con­trol egg ex­tracts and sperm (open bar). Sec­ond graph dis­plays mean per­cent­age of formed nu­clei with γ-tubu­lin lo­cal­ized through­out the nu­clei (black bar) or mar­gin­al­ized to the nu­clear en­ve­lope (open bar) (± s.d., n = 3). In (A-C, E, F) γ-tubu­lin and cen­trin are shown as green, α-tubu­lin and NPC as red and nu­clei as blue (DAPI). In (D) Xgrip109 and γTub are shown as green and red, re­spec­tively. Ar­rows and ar­row­heads in­di­cate the lo­ca­tion of cen­tro­somes and the γ-tubu­lin nu­clear bound­ary, re­spec­tively. Scale bars, 10 μm.
alt-text: Fig. 2
2.2. Nuclear membranes and lamin B3 are recruited to γ-tubulin enriched regions
To de­scribe the ex­act func­tion of γ-tubu­lin dur­ing nu­clear as­sem­bly, we mon­i­tored for­ma­tion of the lam­ina and of the NE by im­munos­tain­ing lamin B3, the most abun­dant lamin pre­sent in X. lae­vis eggs (Lourim et al., 1996), and by stain­ing mem­branes with the lipophilic dye Nile Red (Cox and Leno, 1990; Leno and Laskey, 1991; Lu et al., 1997), as well as an­a­lyz­ing γ-tubu­lin lo­cal­iza­tion. De­mem­branated sperm con­tained a bound­ary of γ-tubu­lin but lacked an in­tact lam­ina and NE (Fig. 4B, C; 0 min). Ad­di­tion of egg ex­tract re­cruited lamin B3 and nu­clear mem­branes (NM) to γ-tubu­lin en­riched ar­eas (Fig. 4B, C; 90 min), which sug­gest that lamin B3 and NM may be re­cruited to the NE in a γ-tubu­lin de­pen­dent man­ner.
2.3. The C-terminal region of γ-tubulin regulates nuclear formation
To elu­ci­date the role of γ-tubu­lin, we mon­i­tored nu­clear for­ma­tion in the ab­sence of γ-tubu­lin and found that the chro­matin was nei­ther able to form lam­ina or to re­cruit NM (Fig. 5A). Ad­di­tion of hu­man γ-tubu­lin to the Ksperm en­abled the for­ma­tion of the NE and of the lam­ina mesh­work again (Fig. 5A).
To iden­tify the γ-tubu­lin-do­main im­por­tant in nu­clear for­ma­tion, we tested var­i­ous γ-tubu­lin mu­tants (Fig. 5B). Bac­te­ri­ally pro­duced His6-γtubu­lin and His6-C-γtub334452, but not His6-N-γtub1333 bound to chro­matin dur­ing nu­clear for­ma­tion (Fig. 5B) and abol­ished the in­hibit­ing ef­fect of γ-tubu­lin de­ple­tion (Fig. 5A). This ob­ser­va­tion proves that both ex­oge­nous γ-tubu­lin and its C-ter­mi­nal do­main bind back to chro­matin. In ad­di­tion, it also demon­strates that the C-ter­mi­nal do­main is the nec­es­sary do­main for nu­clear and lam­ina as­sem­bly and no­tably, this do­main con­tains γ-tubu­lins DNA bind­ing mo­tif (Fig. 1A) (Hoog et al., 2011).
2.4. γ-Tubulin recruitment of nuclear membranes is independent of lamina formation
In­ter­fer­ence with the func­tion of lamin B3 af­fects the size of the formed nu­clei (Lourim et al., 1996). To test whether the de­fects in nu­clear as­sem­bly ob­served fol­low­ing γ-tubu­lin de­ple­tion are due to an im­paired lam­ina for­ma­tion or lamin B3 re­cruit­ment to chro­matin, we added re­com­bi­nant lamin B3 to Ksperm (Fig. 6A). In the ab­sence of chro­matin-bound γ-tubu­lin, lamin B3 did not as­so­ci­ate with Ksperm and, con­se­quently, ad­di­tion of re­com­bi­nant lamin B3 was not suf­fi­cient to trig­ger nei­ther lam­ina nor nu­clear for­ma­tion (Fig. 6A),
UNCORRECTED PROOF
Fig. 3. Nu­clear γ-tubu­lin is nec­es­sary for nu­clear as­sem­bly. (A) Nu­clear as­sem­bly was per­formed as in Fig. 2B. Egg ex­tracts (egg extr.) were im­mun­ode­pleted in the pres­ence (Depl. γTub) or ab­sence (Depl. Cont.) of an anti-γ-tubu­lin an­ti­body by im­muno­pre­cip­i­ta­tion and pro­tein lev­els of γ-tubu­lin and α-tubu­lin were an­a­lyzed by WB (n = 6). (B) Graph shows the mean per­cent­age of formed nu­clei in stage 3 and 4 rel­a­tive to a con­trol (black bar) in nu­clear as­sem­bly re­ac­tions that were per­formed un­der the con­di­tions de­scribed in D (± s.d., n = 3, ** p < 0.01). (C) Pro­teins as­so­ci­ated with the de­mem­branated sperm were de­graded with pro­teinase K (ptK; Ksperm), as in­di­cated and re­main­ing pro­tein lev­els of γ-tubu­lin and hi­s­tone 2B were an­a­lyzed by west­ern blot­ting (WB). The lo­cal­iza­tion of γ-tubu­lin and chro­matin in sperm were as­sayed by im­muno­flu­o­res­cence with an anti-γ-tubu­lin an­ti­body (green; ab27074) and by DAPI stain­ing (blue; n = 3). (D) Schematic rep­re­sen­ta­tion de­scrib­ing the con­di­tions used dur­ing nu­clear as­sem­bly as­says. The fig­ure shows rep­re­sen­ta­tive im­ages from at least ten ex­per­i­ments. Scale bars, 10 μm.
alt-text: Fig. 3
sug­gest­ing that lamin B3 needs to be re­cruited to chro­matin by γ-tubu­lin.
Fi­nally, to demon­strate that the role of γ-tubu­lin in nu­clear for­ma­tion is in­de­pen­dent of the for­ma­tion of the lam­ina, we im­mun­ode­pleted lamin B3 from egg ex­tracts and stud­ied the lo­ca­tion of γ-tubu­lin and NM. In­deed, the lack of lamin B3 nei­ther af­fected the lo­ca­tion of γ-tubu­lin or the re­cruit­ment of NM (Fig. 6B). These data prove that re­cruit­ment of NM by the nu­clear γ-tubu­lin bound­ary is in­de­pen­dent from for­ma­tion of the lam­ina mesh­work.
2.5. Endogenous γ-tubulin forms a cellular protein meshwork
In an asyn­chro­nous cell pop­u­la­tion, ap­prox­i­mately 26% of the to­tal amount of en­doge­nous γ-tubu­lin is as­so­ci­ated with chro­matin in U2OS and NI­H3T3 cells (Eklund et al., 2014; Hoog et al., 2011). To un­der­stand the in­ter­con­nec­tion be­tween the cy­toso­lic and the nu­clear γ-tubu­lin pools, we per­formed con­fo­cal mi­croscopy and su­per­res­o­lu­tion mi­croscopy of the top plane (Fig. 7A) and mid plane (Fig. 7B and Fig. 8) of fixed U2OS cells and of liv­ing U2OS cells that sta­bly co-ex­pressed γTUBU­LIN shRNA (γTUBU­LINsh-U2OS) and hu­man GFP-tagged sh-re­sis­tant γ-tubu­lin (γTUBU­LINsh-U2OS-GFP-γ-tubu­linresist; Fig. 9A, B) and tran­siently ex­pressed mCherry-tagged lamin B (mCherry-lamin B1; Fig. 9B). The γTUBU­LIN shRNA re­duced the ex­pres­sion of en­doge­nous γ-tubu­lin by ap­prox­i­mately 50% (53 ± 6%, n = 3; Fig. 9A) and we com­pen­sated for this re­duc­tion by sta­bly co-ex­press­ing GFP-γ-tubu­linresist (Fig. 9A, B).
Con­fo­cal mi­croscopy of the top plane (Fig. 7A) and mid plane (Fig. 7B and Fig. 8) of fixed U2OS cells and Z-stack im­ages of whole liv­ing γTUBU­LINsh-U2OS-GFP-γ-tubu­linresist cells (Fig. 9B) showed that on the NE, en­doge­nous and re­com­bi­nant γ-tubu­lin formed strings and to­gether with lamin B formed an in­ter­con­nected pro­tein mesh
­work on both the lower (Fig. 9B; 0.341.36 μm) and on the higher nu­clear sur­face (Fig. 7A and Fig. 9B; 3.063.74 μm). On the NE, en­doge­nous γ-tubu­lin strings looked sim­i­lar to lamin B fibers (Fig. 7A). How­ever, the lamin B1 mesh­work be­came non-de­tectable in the nu­clear com­part­ment (Fig. 9B; 0.683.06 μm). In con­trast, the γ-tubu­lin strings were de­tected through­out the nu­cleus (Fig. 9B; 0.343.74 μm) and es­tab­lished γ-tubu­lin pro­tein bridges be­tween the cy­to­plasm and the nu­clear com­part­ment (Fig. 7B, Fig. 8 and Fig. 9B).
To ex­clude the pos­si­bil­ity that fix­a­tion of the cells caused for­ma­tion of γ-tubu­lin strings, we stud­ied the lo­ca­tion of the pro­tein cen­trin. In com­par­i­son with γ-tubu­lin, im­muno­flu­o­res­cence stain­ing with an anti-cen­trin an­ti­body showed that cen­trin formed no cel­lu­lar strings (Fig. 9C, D). Ac­cord­ingly, the lo­cal­iza­tion of both the nu­clear GFP-tagged hu­man RNA-bind­ing pro­tein 3 (GFP-RBM3) and of the cy­toso­lic GFP-tagged ser/​thr ki­nase SADB-long (mSADB, hSAD1/​BRSK1; GFP-SADBL) dif­fered from γ-tubu­lin in the NE. Z-stack im­ages of whole liv­ing U2OS cells tran­siently ex­press­ing ei­ther GFP-RBM3 or GFP-SADBL showed that nei­ther GFP-RBM3 nor GFP-SADBL were found to con­tinue cross the NE (0.340.68 μm and 2.723.74 μm RBM3 and 0.340.68 μm and 3.063.74 μm SADBL; Fig. 7A, B, Fig. 9B and Fig. 10A, B). Al­to­gether, these data con­firm that γ-tubu­lin forms a mesh­work that con­nects the chro­matin to the cy­to­plasm.
2.6. The γ-tubulin meshwork consists of strings
To char­ac­ter­ize the struc­ture of the γ-tubu­lin mesh­work, we per­formed im­mu­no­elec­tron mi­croscopy of U2OS cells that were pre­pared with two dif­fer­ent meth­ods. In the first method, we high-pres­sure frozen U2OS cells (Fig. 11). In the sec­ond method, we tested var­i­ous fix­a­tion pro­ce­dures and found that short fix­a­tion (5 min) of
UNCORRECTED PROOF
Fig. 4. Re­com­bi­nant γ-tubu­lin re­stores nu­clear as­sem­bly. (A) Nu­clear as­sem­bly was per­formed as in Fig. 3. Bac­te­ri­ally pro­duced His6γ-tubu­lin (HisγTub) or His6E2F1Δ194426 (HisE2FΔ194426) was added back (ad­dback) to Ksperm or to im­mun­ode­pleted egg ex­tract (egg extr.) and the ef­fects of His6γ-tubu­lin on nu­clear as­sem­bly were ex­am­ined. The pro­tein lev­els of γ-tubu­lin in ex­tracts (egg extr.-Ksperm) or sperm were an­a­lyzed by WB as in­di­cated. An α-tubu­lin (αTub) load­ing con­trol is shown (n = 3-6). Graphs show mean per­cent­age of formed nu­clei in stage 3 and 4 rel­a­tive to a con­trol (black bar) from nu­clear as­sem­bly re­ac­tions that con­sisted of im­mun­ode­pleted egg ex­tracts and sperm or Ksperm, in which an anti-γ-tubu­lin an­ti­body (open bar; T3320) or an anti-γ-tubu­lin an­ti­body that af­ter im­mun­ode­ple­tion, ei­ther His6γ-tubu­lin or His6E2F1Δ194426 was added to Ksperm or to de­pleted egg ex­tract (grey bar; ± s.d., n = 3-6 for each graph; * p < 0.05, ** p < 0.01). (B, C) Con­fo­cal im­ages of the mor­pho­log­i­cal changes of nu­clei in stage 1, 3 and 4 from a nu­clear as­sem­bly of egg ex­tracts and sperm treated as in Fig. 3 but in­cu­bated for 90 min be­fore fix­a­tion to in­crease the num­ber of nu­clei in stage 4. Ar­row­heads show γ-tubu­lin bound­aries around sperm and nu­clei. Ar­rows show γ-tubu­linlamin B3 en­riched ar­eas. In (A-C) γ-tubu­lin (γTub; green;) and lamin B3 (lam­inB; red) are shown as im­muno­flu­o­res­cence stain­ing with an anti-γ-tubu­lin and an anti-lamin B3 an­ti­body. Nu­clear mem­branes and nu­clei were de­tected with Nile red (red) and DAPI (blue), re­spec­tively. The fig­ure shows rep­re­sen­ta­tive im­ages from at least ten ex­per­i­ments. Scale bars, 10 μm.
alt-text: Fig. 4
cells with 4% paraformalde­hyde pre­served the γ-tubu­lin mesh­work. With both meth­ods, we de­tected strings in both cy­to­plasm and nu­cleus that went across the NE (Fig. 11A and Fig. 12A). Im­mu­no­elec­tron mi­croscopy con­firmed that the an­ti­body rec­og­nized strings with a 4 to 6 nm in di­am­e­ter, which from now on will be re­ferred as γ-strings (Fig. 11A and Fig. 12A; n = 19). γ-Strings were at­tached to the plasma mem­brane (Fig. 11A and Fig. 12A), oc­curred in both the outer, the in­ner nu­clear mem­brane (Fig. 11A and Fig. 12A) and in the nu­clear com­part­ment (Fig. 11A and Fig. 12A) and con­nected both the nu­clear and the cy­toso­lic γ-tubu­lin pools across the NE (Fig. 11A and Fig. 12A). By con­trast, con­trol im­munos­tain­ing with an anti-α-tubu­lin an­ti­body rec­og­nized cy­toso­lic ar­rays of mi­cro­tubules (Fig. 11B). To­gether these data demon­strate the ex­is­tence of a NE-as­so­ci­ated net­work, the γ-string mesh­work.
To fi­nally prove that the ob­served γ-strings are made of γ-tubu­lin and to study the in vitro ef­fect of γ-tubu­lin on lam­ina for­ma­tion, we in­ves­ti­gated the in vitro abil­ity of γ-tubu­lin to form strings and to as­sist lam­ina for­ma­tion. Elec­tron mi­croscopy analy­sis of bac­te­ri­ally pro­duced γ-tubu­lin showed that in vitro γ-tubu­lin formed a mesh­work of strings only in the ab­sence of GTP (Fig. 12B). Fur­ther­more, the formed mesh­work sup­ported for­ma­tion of lamin B3 protofil­a­ments (Fig. 12B), con­firm­ing that γ-tubu­lin forms strings that as­sist ini­tial nu­cle­ation of lamin B3 into a protofil­a­ment.
2.7. The protein levels of γ-tubulin affect the integrity of the lamina
A chro­matin-as­so­ci­ated pro­tein mesh­work that fa­cil­i­tates lam­ina for­ma­tion may pro­vide a cell with a scaf­fold that main­tains the NE. To test this, we iso­lated nu­clei from both U2OS, γTUBU­LINsh-U2OS-GFP-γ-tubu­linresist and γTUBU­LINsh-U2OS cells (Fig. 13A-C) (Mendez and Stillman, 2000) and found that the nu­clei con­tained a nu­clear bound­ary of γ-strings to which mi­cro­tubule com­po­nents were as­so­ci­ated on the cy­toso­lic side (Fig. 13A). Iso­lated nu­clei from cy­tocha­lasin B and col­cemid treated cells (Alvarado-Kristensson et al., 2009) had no at­tached mi­cro­tubules (Fig. 13B), but still con­tained a γ-string bound­ary in­ter­twined with the lam­ina (Fig. 13C). In ad­di­tion, iso­lated nu­clei from γTUBU­LINsh-U2OS cells showed that 86% of cells with low ex­pres­sion of γ-tubu­lin had a scat­tered lamin B mesh­work (Fig. 13D), which im­ply that the γ-tubu­lin-nu­clear bound­ary formed at the tran­si­tion be­tween cy­toso­lic and chro­matin-as­so­ci­ated γ-strings may func­tion as a sup­port­ing scaf­fold for the lam­ina.
Fi­nally, to ex­am­ine the in­ter­ac­tions of γ-strings with the lam­ina, we an­a­lyzed the lamin and the tubu­lin con­tent of en­doge­nous γ-tubu­lin and lamin B im­muno­pre­cip­i­tated from cy­to­plas­mic, nu­clear mem­brane and chro­matin frac­tions of NI­H3T3 and U2OS cells. We found that al­though as­so­ci­a­tion with the other lam­ina com­po­nents, lamin A and C, was only ob­served in the chro­matin frac­tions, γ-tubu­linlamin
UNCORRECTED PROOF
Fig. 5. Nu­clear-γ-tubu­lin pro­motes the for­ma­tion of the nu­clear en­ve­lope and the lam­ina mesh­work. (A, B) Ksperm were in­cu­bated in the pres­ence of γ-tubu­lin im­mun­ode­pleted egg ex­tracts (Depl.) dur­ing 90 min be­fore fix­a­tion or be­fore spun down through a su­crose cush­ion for analy­sis by WB (ass. nu­clei). Rep­re­sen­ta­tive con­fo­cal flu­o­res­cence im­ages of mor­pho­log­i­cal changes of nu­clei in stage 1 or 4 show­ing the lo­ca­tion of en­doge­nous γ-tubu­lin (eγTub, green) or bac­te­ri­ally pro­duced His6γ-tubu­lin (γTub, green), His6C-γtub334452 (CγTub, green) or His6N-γtub1333 (NγTub, green) of nu­clei in stage 1 or 4. Lo­cal­iza­tion of lamin B3 (lamin B; red) were ex­am­ined by im­muno­flu­o­res­cence stain­ing with an anti-lamin B3 an­ti­body and nu­clear mem­branes and nu­clei were de­tected with Nile red (red) and DAPI (blue), re­spec­tively. Lo­cal­iza­tion of γ-tubu­lin, His6γ-tubu­lin, His6C-γtub334452 and His6N-γtub1333 were im­muno­flu­o­res­cence stained with ei­ther T3320 (eγTub, γTub, CγTub) or T5192 (NγTub). The fig­ure shows rep­re­sen­ta­tive im­ages from at least five ex­per­i­ments. Scale bars, 10 μm. (B) Graph shows the mean per­cent­age of as­sem­bled nu­clei in stage 3 and 4 ver­sus a con­trol (black bar), with an anti-γ-tubu­lin an­ti­body (open and grey bars) and with each form of His6γ-tubu­lin (grey bars) added back (n = 3; * p < 0.05), as in­di­cated. The amount of pro­tein added back (in­put), the γ-tubu­lin level pre­sent in the ex­tracts and the as­so­ci­a­tion of His6γ-tubu­lin, His6C-γtub334452 and His6N-γtub1333 with as­sem­bled nu­clei were an­a­lyzed by WB with the in­di­cated an­ti­bod­ies. Num­bers on WBs in­di­cate the level of de­ple­tion of γ-tubu­lin in the ex­tracts rel­a­tive to con­trol. To ad­just for dif­fer­ences in pro­tein load­ing, the pro­tein con­cen­tra­tion of γ-tubu­lin was de­ter­mined by its ra­tio with α-tubu­lin for each sam­ple. The pro­tein ra­tio in con­trol ex­tracts was set to 1.
alt-text: Fig. 5
Fig. 6. γ-Tubu­lin pro­motes the for­ma­tion of the nu­clear en­ve­lope in the ab­sence of lamin B3. (A) His6lamin B3 (lam­inB3) was added back to Ksperm (Cont.; Ksperm with­out lam­inB3) and γ-tubu­lin im­mun­ode­pleted egg ex­tract (Depl.) and nu­clear as­sem­bly was per­formed as in Fig. 5. Graph shows the mean per­cent­age of formed nu­clei in stage 3 and 4 in im­mun­ode­pleted egg ex­tracts and sperm (Depl. γTub) ver­sus a con­trol (black bar; Depl. Cont.), with ad­di­tion of His6lamin B3 (grey bar) to the γ-tubu­lin de­pleted ex­tracts (± s.d., n = 3; * p < 0.05), as in­di­cated. The west­ern blot shows the amount of His6lamin B3 added to sperm. To re­late the amount of His6lamin B3 to the amount pre­sent in ex­tracts, 1 μl and 5 μl of egg ex­tracts were loaded. Rep­re­sen­ta­tive con­fo­cal flu­o­res­cence im­ages of mor­pho­log­i­cal changes of nu­clei show the lo­ca­tion of en­doge­nous γ-tubu­lin (eγTub, green) or lamin B (red). (B) Rep­re­sen­ta­tive con­fo­cal flu­o­res­cence im­ages of mor­pho­log­i­cal changes of nu­clei in nu­clear as­sem­bly re­ac­tions that were trig­gered by ad­di­tion of lamin B3 im­mun­ode­pleted egg ex­tracts to sperm and in­cu­bated 90 min be­fore fix­a­tion. The pro­tein lev­els of lamin B3 in ex­tracts were an­a­lyzed by WB. The graph shows the mean per­cent­age of formed nu­clei with γ-tubu­lin lo­cal­ized through­out the nu­clei (black bar) or mar­gin­al­ized to the nu­clear en­ve­lope (open bar) (± s.d., n = 3). (A, B) Lo­cal­iza­tion of γ-tubu­lin (eγTub; green), His6lamin B3 (lamin B; red) and lamin B3 (lamin B; red) were ex­am­ined by im­muno­flu­o­res­cence stain­ing with the in­di­cated an­ti­body and nu­clear mem­branes and nu­clei were de­tected with Nile red (red) and DAPI (blue), re­spec­tively. The fig­ure shows rep­re­sen­ta­tive im­ages from at least five ex­per­i­ments. Scale bars, 10 μm.
alt-text: Fig. 6
UNCORRECTED PROOF
Fig. 7. En­doge­nous γ-tubu­lin bridges con­nect both the nu­clear and the cy­toso­lic γ-tubu­lin pools across the nu­clear en­ve­lope. (A, B) Lo­cal­iza­tion of en­doge­nous γ-tubu­lin with an anti-γ-tubu­lin (T3320; green), lam­ina with an anti-lamin B (lam­inB; red) an­ti­body and nu­clei with DAPI (blue) were ex­am­ined in U2OS cells in in­ter­phase. Con­fo­cal im­ages are the mid planes of the γ-tubu­lin bound­ary at the nu­clear mem­brane (A) or of the nu­clear com­part­ment (B) a U2OS cell. The yel­low box shows colo­cal­iza­tion pixel-map (CM) of the red and green chan­nels of the mag­ni­fied ar­eas dis­played in the in­set. White ar­eas in CM de­note colo­cal­ized pix­els be­tween chan­nels. The fig­ure shows rep­re­sen­ta­tive im­ages from at least five ex­per­i­ments. Scale bars, 10 μm.
alt-text: Fig. 7
B com­plexes were im­muno­pre­cip­i­tated from all cel­lu­lar frac­tions (Fig. 13E; Fig. S1), sug­gest­ing a func­tion of γ-tubu­lin in lamin B re­cruit­ment to the NE.
2.8. The γ-string boundary supports the formation of the lamina
Based on the find­ing of the γ-string bound­ary, we hy­poth­e­sized that it may pro­vide a cell with a tool to struc­ture syn­chro­nized cy­toso­lic and nu­clear events dur­ing cell di­vi­sion. To vi­su­al­ize the in­ter­con­nec­tion be­tween γ-strings and lam­ina for­ma­tion dur­ing nu­clear
as­sem­bly, we an­a­lyzed time-lapse im­ages of di­vid­ing γTUBU­LINsh-U2OS­GFP-γ-tubu­linresist (with C-ter­mi­nal tagged γ-tubu­lin; Fig. 14; movie S1), U2OS (Fig. 15A; movie S2) and γTUBU­LINsh-U2OS cells (Fig. 15B; movie S3) that tran­siently co-ex­pressed mCher­ry­lamin B1. In metaphase, γ-tubu­lin was evenly dis­trib­uted through­out the cell (Fig. 14, 0 min) and a γ-tubu­lin-bound­ary com­posed of cy­toso­lic and chro­matin-as­so­ci­ated γ-strings around the mi­totic chro­mo­somes (Fig. 14, 03 min) was ob­served. This γ-tubu­lin-bound­ary re­mained in the newly formed daugh­ter nu­clei (Fig. 14, 412 min). In con­trast, in metaphase lamin B re­mained non-chro­matin bound (Fig. 14, 03 min). Dur­ing early telophase the lam­ina be­came vis­i­ble and grew over time in­ter­twined with the γ-tubu­lin-bound­ary of γ-strings (Fig. 14, 412 min). Struc­tured il­lu­mi­na­tion mi­croscopy con­firmed the lo­ca­tion of the γ-tubu­lin-bound­ary in the newly as­sem­bled daugh­ter nu­clei (Fig. 14), which re­sem­bled the γ-tubu­lin nu­clear bound­ary formed in stage 4 in a newly as­sem­bled X. lae­vis nu­cleus (Fig. 4B). Sim­i­lar mCherry-lamin B1 lo­cal­iza­tion was ob­served in U2OS cells (Fig. 15A) and in γTUBU­LINsh-U2OS cells, de­spite the lower ex­pres­sion of en­doge­nous γ-tubu­lin and mCherry-lamin B1 in the lat­ter cell line (Fig. 15B). These data fur­ther sup­port that in liv­ing cells γ-strings may func­tion as a scaf­fold mesh­work that as­sists the for­ma­tion of the lam­ina around mi­totic chro­mo­somes.
As the amount of GFP-γ-tubu­lin as­so­ci­ated with mi­totic chro­mo­somes was higher than an­tic­i­pated in γTUBU­LINsh-U2OS-GFP-γ-tubu­linresist cells (Fig. 14), we tested the ef­fect of the po­si­tion of the GFP-tag and the sta­ble ex­pres­sion of γTUBU­LIN shRNA on the cel­lu­lar lo­ca­tion of GFP-γ-tubu­lin in γtubGFP-U2OS (U2OS cells that sta­bly ex­pressed hu­man GFP-γ-tubu­lin; Fig. 16A) and in γTUBU­LINsh-U2OS-NGFP-γ-tubu­linresist cells (γTUBU­LINsh-U2OS cells that sta­bly ex­pressed hu­man GFP-N ter­mi­nal-tagged sh-re­sis­tant γ-tubu­lin; Fig. 16B) and mon­i­tored the amount of chro­matin-as­so­ci­ated
Fig. 8. γ-strings are in­ter­twined with the lam­ina mesh­work. Lo­cal­iza­tion of en­doge­nous γ-tubu­lin with an anti-γ-tubu­lin (T6557; green), lam­ina with an anti-lamin B (lam­inB; red) an­ti­body and nu­clei with DAPI (blue) were ex­am­ined in U2OS cells in in­ter­phase. Struc­tured il­lu­mi­na­tion im­ages are the mid planes of the nu­clear com­part­ment of a U2OS cell. Z-stack im­ages were col­lected at 0.12 μm in­ter­vals. Z-stack shows av­er­age in­ten­sity pro­jec­tion of the dis­played three Z-stack im­ages. The white box shows the mag­ni­fied ar­eas. Ar­row­heads show pro­tein bridges of γ-tubu­lin. The yel­low box shows colo­cal­iza­tion pixel-map (CM) of the red and green chan­nels. White ar­eas in CM de­note colo­cal­ized pix­els be­tween chan­nels. The fig­ure shows rep­re­sen­ta­tive im­ages from four ex­per­i­ments. Scale bars, 10 μm. Right panel, γ-tubu­lin forms strings (γstrings) that act as pro­tein bridges (γTB) be­tween the cy­toso­lic and the nu­clear com­part­ment. In their way into the nu­cleus, a γ-string goes through the nu­clear lam­ina.
alt-text: Fig. 8
UNCORRECTED PROOF
Fig. 9. γ-Tubu­lin forms a cel­lu­lar mesh­work. (A) To­tal lysate from U2OS, U2OS cells sta­bly ex­press­ing γTUBU­LIN shRNA (shγTUB) and U2OS cells sta­bly ex­press­ing γTUBU­LIN shRNA and GFP-γ-tubu­linresist (γTubresist, n = 3) were an­a­lyzed by WB for the ex­pres­sion of en­doge­nous γ-tubu­lin (ar­row­head) and GFP-γ-tubu­linresist (ar­row) with a mix­ture of an anti-γ-tubu­lin (T6557) and an anti-GFP (sc-8334) an­ti­body. An α-tubu­lin load­ing con­trol is shown. (B) Dif­fer­en­tial in­ter­fer­ence con­trast (DIC)/​flu­o­res­cence im­ages of a Z-stack of U2OS cells sta­bly ex­press­ing γTUBU­LIN shRNA and GFP-γ-tubu­linresist (γ-TubGFP; green) and tran­siently ex­press­ing mCherry-tagged lamin B (red). Im­ages were col­lected at 0.34 μm in­ter­vals. Z-stack im­ages taken in the in­ter­vals 0.341.36 μm and 3.063.74 μm show the lower and the higher γ-string mesh­work formed on the nu­clear en­ve­lope, re­spec­tively. Z-stack im­ages in the in­ter­val 1.72.72 μm show the γ-string mesh­work in the nu­clear com­part­ment. Ar­rows and ar­row­heads show the cen­tro­somes and γ-tubu­lin bridges, re­spec­tively. The mag­ni­fied area (white bor­der) shows a γ-tubu­lin bridge (right pan­els). The fig­ure shows rep­re­sen­ta­tive im­ages from at least eight ex­per­i­ments. (C, D) Lo­cal­iza­tion of en­doge­nous cen­trin with an anti-cen­trin (green) and lam­ina with an anti-lamin B (lam­inB; red) an­ti­body were ex­am­ined in U2OS cells in in­ter­phase. Con­fo­cal im­ages are the mid planes of the nu­clear mem­brane (C) or of the nu­clear com­part­ment (D) of U2OS cells. (B-D) Scale bars, 10 μm.
alt-text: Fig. 9
Fig. 10. Nei­ther RBM3 nor SadBL form a cel­lu­lar mesh­work. (A, B) Con­fo­cal im­ages of a Z-stack of U2OS cells tran­siently ex­press­ing ei­ther GFP-RBM3 (RBM3; A) or GFP-SadBL (SadB; B), and tran­siently ex­press­ing mCherry-tagged lamin B (lam­inB1). Im­ages were col­lected at 0.34 μm in­ter­vals. (A) Z-stack im­ages taken in the in­ter­vals 0.340.68 μm and 2.723.74 μm show the nu­clear en­ve­lope and in the in­ter­val 1.022.68 μm shows the nu­clear com­part­ment of cells ex­press­ing RBM3. (B) Z-stack im­ages taken in the in­ter­vals 0.340.68 μm and 3.063.74 μm show the nu­clear en­ve­lope and in the in­ter­val 1.022.72 μm shows the nu­clear com­part­ment of cells ex­press­ing SADB. (A, B) The mag­ni­fied area (white bor­der) shows the mid plane of the nu­clear com­part­ment (right pan­els). Scale bars, 10 μm.
alt-text: Fig. 10
UNCORRECTED PROOF
Fig. 11. The cy­toso­lic and the nu­clear pools of γ-tubu­lin are con­nected and form one pro­tein mesh­work. (A, B) En­doge­nous γ-tubu­lin de­tec­tion in im­mu­no­elec­tron mi­croscopy us­ing three dif­fer­ent con­di­tions in high pres­sure frozen (HPF) U2OS cells: first, no an­ti­body (A), sec­ond, gold con­ju­gated pro­tein A (A) and third, an anti-γ-tubu­lin an­ti­body (T3320; A) and gold con­ju­gated pro­tein A (RγTub; A) or an anti-α-tubu­lin an­ti­body and gold con­ju­gated pro­tein A (RαTub; B). Im­ages show the plasma mem­brane [PM], cy­tosol [C], in­ner [IN] and outer [ON] nu­clear mem­branes and nu­cleus [N] and γ-tubu­lin bridges [γTB] of a U2OS cell. A black box shows the mag­ni­fied area dis­played in the in­set. Black ar­row­heads show ei­ther im­muno­la­beled γ-strings (RγTub) or im­muno­la­beled mi­cro­tubules (RαTub). Ar­rows show the in­di­cated struc­ture (n = 3).
alt-text: Fig. 11
Fig. 12. The γ-tubu­lin mesh­work sup­ports for­ma­tion of lamin B3 protofil­a­ments. (A) En­doge­nous γ-tubu­lin de­tec­tion in im­mu­no­elec­tron mi­croscopy us­ing an anti-γ-tubu­lin an­ti­body or no pri­mary an­ti­body (con­trol) on paraformalde­hyde fixed (PFA) U2OS cells. Im­ages show the plasma mem­brane [PM], cy­tosol [C], in­ner [IN] and outer [ON] nu­clear mem­branes and nu­cleus [N] and γ-tubu­lin bridges [γTB] of a U2OS cell. A black box shows the mag­ni­fied area dis­played in the in­set. Black ar­row­heads show ei­ther im­muno­la­beled γ-strings (RγTub) or im­muno­la­beled mi­cro­tubules (RαTub). Ar­rows show the in­di­cated struc­ture (n = 3). (B) Pu­ri­fied γ-tubu­lin and lamin B3 (lamB3) were neg­a­tively stained and im­aged by elec­tron mi­croscopy in the ab­sence (γ-Tub) or pres­ence of 1 mM GTP (γ-TubGTP). The mag­ni­fied area (black bor­der) shows γ-strings and ar­rows and ar­row­heads show lamin B3 fibers and γ-strings, re­spec­tively.
alt-text: Fig. 12
GFP-γ-tubu­lin pool by time-lapse mi­croscopy. We no­ticed that the as­so­ci­a­tion of GFP-γ-tubu­lin with chro­matin was in­de­pen­dent of the po­si­tion of the GFP-tag. Also, the lower the en­doge­nous pro­tein lev­els of γ-tubu­lin, the higher lev­els of chro­matin-as­so­ci­ated GFP-γ-tubu­lin were found (Fig. 14 and Fig. 16C). These find­ings sug­gested that en­doge­nous γ-tubu­lin in­ter­fered with GFP-γ-tubu­lin bind­ing to chro­matin and that the amount of chro­matin-as­so­ci­ated γ-tubu­lin might be higher than ex­pected.
2.9. Endogenous γ-tubulin is associated to mitotic chromosomes
One plau­si­ble rea­son for the un­der­es­ti­ma­tion of the amount of chro­matin-as­so­ci­ated γ-tubu­lin is that avail­able an­ti­bod­ies may not fully rec­og­nize this pool, as pre­vi­ously re­ported (Eklund et al., 2014). In­deed, in­ter­phase (Fig. 17A) and mi­totic (Fig. 17B) γ-tubu­lin im­munos­tained γTUBU­LINsh-U2OS-GFP-γ-tubu­linresist cells showed that the an­ti­body rec­og­nized only part of the chro­matin-as­so­ci­ated pool (Fig. 17A, B). Nonethe­less, struc­tured il­lu­mi­na­tion mi­croscopy
UNCORRECTED PROOF
Fig. 13. γ-Tubu­lin forms a bound­ary on the cy­toso­lic side of the nu­cleus and its dis­rup­tion af­fects the in­tegrity of the lam­ina. (A-D) Con­fo­cal im­ages of iso­lated nu­clei pu­ri­fied in the ab­sence (A) or pres­ence (B-D) of cy­tocha­lasin B and col­cemid from U2OS (γTub) and sta­ble γTUBU­LIN shRNA-GFP-γ-tubu­linresist (GFPγTub) and γTUBU­LIN shRNA ex­press­ing cells. (A-C) White bor­ders show the mag­ni­fied ar­eas dis­played in the in­sets: the γ-string mesh­work on the nu­clear en­ve­lope (A-C), the γ-string bound­ary (A, B), chro­matin-as­so­ci­ated γ-strings (B) or γ-string bridges (C). In (A-D) γ-tubu­lin (γTub; green, T3320), α-tubu­lin (αTub; red) and lamin B (lam­inB; red) are shown as im­muo­flu­o­res­cence stain­ing and nu­clei were de­tected with DAPI (blue). (D) shows the en­doge­nous ex­pres­sion of lamin B and γ-tubu­lin in two nu­clei con­tain­ing ei­ther high or low γ-tubu­lin ex­pres­sion, as in­di­cated. (A-D) Scale bars, 10 μm. (E) U2OS and NI­H3T3 cells (20 × 106 cells) were bio­chem­i­cally di­vided into cy­toso­lic [C], nu­clear mem­brane [N], and chro­matin [CH] frac­tions. Each frac­tion was sub­jected to im­muno­pre­cip­i­ta­tions (IP) with an anti-γ-tubu­lin (γTub; T6557), anti-lamin B or anti-GFP (Cont.) an­ti­body, as in­di­cated, and de­vel­oped by WB with an­ti­bod­ies against lamin A/​C, lamin B, γ-tubu­lin (T5192) and α-tubu­lin an­ti­body (n = 5). (A-E) The fig­ure shows rep­re­sen­ta­tive im­ages from at least five ex­per­i­ments. See also Fig. S1.
alt-text: Fig. 13
Fig. 14. Lamin B1 as­sem­bles at the γ-tubu­lin bound­ary. The DIC/​flu­o­res­cence im­ages show time-lapse se­ries from a sta­ble γTUBU­LINsh-U2OS cell co-ex­press­ing GFP-γ-tubu­linresist (γTub; green) that was tran­siently ex­press­ing mCherry-lamin B1 (lamB1; red). The im­age se­ries show cho­sen frames of the lam­ina (lam­inB1) for­ma­tion at the γ-string bound­ary (cy­toso­lic and chro­matin-as­so­ci­ated γ-strings) formed by GFP-γ-tubu­linresist dur­ing nu­clear as­sem­bly in a mi­totic cell. The mi­totic chro­mo­somes/​daugh­ter nu­clei are mag­ni­fied be­low the merged im­ages. Ar­row­heads and ar­rows show the γ-string-bound­ary around the mi­totic chro­mo­somes and the mi­totic spin­dle, re­spec­tively. The graph shows the time-de­pen­dent changes in flu­o­res­cence in­ten­sity across the white box at the γ-string-bound­ary of γtubGFP (green) and mCherry-lamin B1 (mChlmB1; red) ex­pressed in ar­bi­trary units (AU; mean ± s.d.; n = 3). Right, struc­tured il­lu­mi­na­tion mi­cro­scope (SIM) im­ages show im­muno­flu­o­res­cence stain­ing of en­doge­nous γ-tubu­lin and lamin B with an anti-γ-tubu­lin (green; T6557) and anti-lamin B1 (red) an­ti­body of newly formed U2OS daugh­ter cells. Nu­clei were de­tected with DAPI (blue). The yel­low box shows colo­cal­iza­tion pixel-map (CM) of the red and green chan­nels. White ar­eas in CM de­note colo­cal­ized pix­els be­tween chan­nels. Scale bars, 10 μm. See also movie S1.
alt-text: Fig. 14
de­tected the chro­matin-as­so­ci­ated γ-tubu­lin pool in mi­totic cells (Fig. 17C), con­firm­ing the as­so­ci­a­tion of γ-tubu­lin with mi­totic chro­mo­somes.
To ob­tain an an­ti­body that bet­ter rec­og­nized chro­matin-as­so­ci­ated γ-tubu­lin, we gen­er­ated a poly­clonal rab­bit an­ti­body (385Ab) to the
γ-tubu­lin re­gion con­tain­ing Ser385 (residues 372 to 389), as phos­pho­ry­la­tion of this residue in­duces a change in the con­for­ma­tion of γ-tubu­lin that causes a size shift from 49 kDa to 60 kDa in SDS gels orig­i­nat­ing a pro­tein band that is rec­og­nized by nei­ther com­mer­cially
UNCORRECTED PROOF
Fig. 15. For­ma­tion of the lam­ina in U2OS and γTUBU­LINsh-U2OS cells. (A, B) The DIC/​flu­o­res­cence im­ages show time-lapse se­ries from U2OS (A) and γTUBU­LINsh-U2OS cells (B; shγTUB) that were tran­siently ex­press­ing mCherry-lamin B1 (lamB1). The im­age se­ries show cho­sen frames of the lo­ca­tion of lamin B1 dur­ing nu­clear as­sem­bly in a mi­totic cell. Im­ages were col­lected every 30 sec. Scale bars, 10 μm. Ar­row­heads show the formed lam­ina around daugh­ter chro­matids. These time-lapse movies are avail­able at movie S2 and movie S3.
alt-text: Fig. 15
Fig. 16. The en­doge­nous pro­tein lev­els of γ-tubu­lin af­fect the as­so­ci­a­tion of GFP-γ-tubu­lin to chro­matin. (A, B) The DIC/​flu­o­res­cence im­ages show time-lapse se­ries from U2OS cells that sta­bly ex­pressed ei­ther C-ter­mi­nal tagged GFP-γ-tubu­lin (γTub; A) or both γTUBU­LIN shRNA and N-ter­mi­nal tagged GFP-γ-tubu­linresist (γTubresistShγTUB; B). The im­age se­ries show cho­sen frames of the vari­a­tion over time in the amount of chro­matin-as­so­ci­ated GFP-γ-tubu­lin. Im­ages were col­lected every 30 sec. Scale bars, 10 μm. (C) The graph shows the time de­pen­dent changes in flu­o­res­cence in­ten­sity across the chro­matin of GFP-γ-tubu­lin ex­pressed in ar­bi­trary units (AU; mean ± s.d.; n = 3).
alt-text: Fig. 16
avail­able anti-γ-tubu­lin nor anti-GFP an­ti­bod­ies (Fig. 17D) (Eklund et al., 2014). The 385Ab rec­og­nized a 60-kDa band, which sig­nal was re­duced upon re­duced pro­tein lev­els of γ-tubu­lin (Fig. 17D), demon­strat­ing that the iden­ti­fied pro­tein band was γ-tubu­lin. Fur­ther­more, the 385Ab rec­og­nized an ad­di­tional 90-kDa band in U2OS cells sta­bly ex­press­ing Nγ-tubGFP that was nei­ther rec­og­nized by anti-γ-tubu­lin or anti-GFP an­ti­bod­ies, but were rec­og­nized by an­ti­bod­ies gen­er­ated to the γ-tubu­lin re­gion con­tain­ing Ser131 (Eklund et al., 2014). Al­to­gether, the data proves that the GFP-tagged γ-tubu­lin and γ-tubu­lin un­dergo the same con­for­ma­tional change (Fig. 17D).
Im­muno­flu­o­res­cence analy­sis showed that in com­par­i­son to other anti-γ-tubu­lin an­ti­bod­ies (Fig. 17A), 385Ab only stained par­tially the γ-tubu­lin pools as­so­ci­ated with cen­tro­somes and mi­cro­tubules (Fig. 18A, B), but rec­og­nized in­stead a γ-tubu­lin pool that was evenly dis­trib­uted through­out in­ter­phase and mi­totic cells (Fig. 18A, B). More­over, the im­muno­flu­o­res­cence stain­ing rec­og­nized with 385Ab was de­creased in U2OS cells ex­press­ing γTUBU­LIN shRNA (Fig. 18B). Fi­nally, im­mu­no­elec­tron mi­croscopy of U2OS cells us­ing 385Ab showed that the an­ti­body fully rec­og­nized γ-tubu­lin bridges (Fig.
18C) and cy­toso­lic and nu­clear γ-strings (Fig. 18D). The data pre­sented here con­firm that dur­ing cell di­vi­sion, there is a γ-tubu­lin bound­ary formed of γ-strings around chro­matin (Fig. 18E).
2.10. Formation of chromatin-containing nuclei depends on the DNA-binding domain of γ-tubulin
The ob­ser­va­tion that the C-ter­mi­nal DNA-bind­ing do­main of γ-tubu­lin (Hoog et al., 2011) is the nec­es­sary do­main for the for­ma­tion of a nu­clear mem­brane and a lam­ina in X. lae­vis egg ex­tracts prompted us to in­ves­ti­gate the role of γ-tubu­lins N- (residues 1333) and C-ter­mi­nal (residues 334452) re­gions in nu­clear for­ma­tion in U2OS cells.
We first in­ves­ti­gated the γ-tubu­lin do­main nec­es­sary in the γ-tubu­linlamin com­plex by an­a­lyz­ing GFP im­muno­pre­cip­i­tates from γTUBU­LINsh-U2OS cells sta­bly ex­press­ing one of the fol­low­ing con­structs: GFP-γ-tubu­linresist, N-ter­mi­nal (NγtubGFP1333; γTUBU­LINsh-U2OS-GFP-Nγ-tubGFP1333) and C-ter­mi­nal (CγtubGFPresist334452; γTUBU­LINsh-U2OS-GFP-Cγ-tubGFP334452)
UNCORRECTED PROOF
Fig. 17. Anti-γ-tubu­lin an­ti­bod­ies par­tially rec­og­nize chro­matin-as­so­ci­ated γ-strings. (A, B) The flu­o­res­cence im­ages show rep­re­sen­ta­tive im­ages of im­munos­tained GFP-γ-tubu­lin (green; γTubGFP) with an anti-γ-tubu­lin an­ti­body (red; γTubAb; T3320) in in­ter­phase (A) and mi­totic (B) U2OS cells sta­bly ex­press­ing both γTUBU­LIN shRNA (shRNA) and GFP-γ-tubu­linresist (γTubGFP). Four dif­fer­ent an­ti­bod­ies were tested (ab27074, T5192, T3320 and T6557) and the one that best rec­og­nized chro­matin-as­so­ci­ated γ-tubu­lin (T3320) is shown. (A) The speci­ficity of the poly­clonal anti-γ-tubu­lin an­ti­body (T3320) used was tested in U2OS cells sta­bly ex­press­ing both γTUBU­LIN shRNA and GFP-γ-tubu­linresist (γTUBU­LINsh-U2OS-GFP-γ-tubu­linresist cells). The dashed line fol­lows one of three γTUBU­LINsh-U2OS-GFP-γ-tubu­linresist cells with low ex­pres­sion of both en­doge­nous γ-tubu­lin and GFP-γ-tubu­linresist. Ar­rows and ar­row­heads in­di­cate a cell with high and low ex­pres­sion of γ-tubu­lin and the lo­ca­tion of cen­tro­somes, re­spec­tively. (C) Struc­tured il­lu­mi­na­tion im­ages of a fixed U2OS im­munos­tained with an anti-γ-tubu­lin (T6557). (A-C) White bor­ders show the mag­ni­fied ar­eas dis­played in the in­sets. Scale bars, 10 μm. (B, C) Nu­clei were de­tected with DAPI (blue). (D) To­tal lysate from U2OS and U2OS cells sta­bly ex­press­ing γTUBU­LIN shRNA (shγTUB) or NGFP-γ-tubu­lin were first an­a­lyzed by WB with a poly­clonal anti-γ-tubu­lin an­ti­body (385Ab) and reprobed with a poly­clonal anti-γ-tubu­lin an­ti­body (131Ab), a mix­ture of two anti-γ-tubu­lin (T5192 and T3320) or of two anti-GFP an­ti­bod­ies (sc-53882 and sc-8334) and an anti-α-tubu­lin an­ti­body. Ar­row­heads in­di­cate the 60-kDa and the 90-kDa bands rec­og­nized by 385Ab (n = 3). Num­bers on WBs in­di­cate the level of de­ple­tion or en­rich­ment of the amount or pro­teins rec­og­nized by the in­di­cated an­ti­body rel­a­tive to con­trol. To ad­just for dif­fer­ences in pro­tein load­ing, the pro­tein con­cen­tra­tion of the in­di­cated pro­teins was de­ter­mined by its ra­tio with en­doge­nous γ-tubu­lin for each sam­ple. The pro­tein ra­tio in con­trol ex­tracts was set to 1.
alt-text: Fig. 17
re­gion of γ-tubu­lin (Fig. 19A). Both CγtubGFP334452 and NγtubGFP1333 were as­so­ci­ated with lamins (Fig. 19A), which im­ply that both re­gions con­tain nec­es­sary se­quences for the for­ma­tion of the γ-tubu­linlamin com­plex. It also sug­gested that the N-ter­mi­nal re­gion of γ-tubu­lin might be suf­fi­cient for trig­ger­ing lam­ina for­ma­tion.
To un­der­stand the im­pact of the in­ter­ac­tions of the N-ter­mi­nal re­gion of γ-tubu­lin with lamin B1 on lam­ina for­ma­tion in U2OS cells, we mon­i­tored mi­totic γTUBU­LINsh-U2OS-GFP-Cγ-tubGFP334452 and γTUBU­LINsh-U2OS-GFP-Nγ-tubGFP1333 (Eklund et al., 2014; Hoog et al., 2011) cells that tran­siently co-ex­pressed mCherry-lamin B1 by time-lapse mi­croscopy. Mi­totic U2OS cells ex­press­ing Cγ-tubGFP334452 di­vided sim­i­larly to γTUBU­LINsh-U2OS-GFP-γ-tubu­linresist cells (Fig. 14 and Fig. 19B; movie S4; n = 16). By con­trast, in 39% (13 cells) of the stud­ied mi­totic γTUBU­LINsh-U2OS-GFP-Nγ-tubGFP1333 cells (n = 33), daugh­ter cells formed an ad­di­tional lam­ina in the ab­sence of chro­matin (movie S6; Fig. 20A, 659 min). More­over, we found chro­matin empty nu­clear like struc­tures in 15% of in­ter­phase cells sta­bly ex­press­ing N-γtubGFP (4 ± 1%, n = 3; Fig. 20B). To­gether, our re­sults demon­strate that both the N- and C-ter­mi­nal re­gions of γ-tubu­lin as­sist in the for­ma­tion of the lam­ina, but only the DNA-bind­ing C ter­mi­nus has the abil­ity to as­sure the for­ma­tion of a lam­ina around chro­matin.
3. Discussion
Here we show that γ-tubu­lin is an im­por­tant co­or­di­na­tor of cy­toso­lic and nu­clear events that leads to the for­ma­tion of the nu­clear mem­brane and the lam­ina in the two daugh­ter cells. γ-Tubu­lin forms cy­toso­lic and chro­matin-as­so­ci­ated γ-strings, which we sug­gest both give sup­port to the emerg­ing nu­clear mem­brane and as­sure the for­ma­tion of a nu­clear en­ve­lope around daugh­ter chro­matids. Im­mun­ode­ple­tion of γ-tubu­lin im­paired nu­clear for­ma­tion in X. lae­vis egg ex­tracts and ad­di­tion of γ-tubu­lin or its C-ter­mi­nal re­gion re­versed this ef­fect. The ef­fect of γ-tubu­lin on nu­clear as­sem­bly is based on γ-tubu­lins abil­ity to di­rectly bind to the sperm chro­matin and to fa­cil­i­tate lam­ina and nu­clear mem­brane for­ma­tion. In mam­malian cell lines, γ-tubu­lin forms γ-strings and the nu­clear en­ve­lope emerged at the γ-tubu­lin bound­ary com­posed of cy­toso­lic and chro­matin as­so­ci­ated γ-strings. In vitro, γ-tubu­lin forms strings that sup­port for­ma­tion of lamin B3 protofil­a­ments and the lam­ina is dis­as­sem­bled in cells with de­creased pro­tein lev­els of γ-tubu­lin. Fi­nally, al­though both γ-tubu­lins N and C ter­mi­nus in­duce the for­ma­tion of a lam­ina, only the DNA-bind­ing do­main of γ-tubu­lin can build up a func­tional nu­cleus (Hoog et al., 2011). Thus, we pro­pose that the in­ter­ac­tion of lamins with the N-ter­mi­nal re­gion of γ-tubu­lin prob­a­bly con­nects laminγ-tubu­lin com­plexes with the cy­toso­lic γ-tubu­lin pool, whereas the C-ter­mi­nal re­gion links laminγ-tubu­lin with chro­matin.
To our knowl­edge we are the first group show­ing the for­ma­tion of a lam­ina in the ab­sence of chro­matin. Stud­ies per­formed with X. lae
UNCORRECTED PROOF
Fig. 18. Anti-γ-tubu­lin an­ti­bod­ies against Ser385-γ-tubu­lin rec­og­nize chro­matin-as­so­ci­ated γ-strings. (A, B) The flu­o­res­cence im­ages show rep­re­sen­ta­tive im­ages of im­munos­tained U2OS and sta­ble γTUBU­LINsh-U2OS cells with 385Ab (green) in in­ter­phase and mi­to­sis. Nu­clei were de­tected with DAPI (blue). Scale bars, 10 μm. (A) Mi­cro­tubules and cen­tro­somes were stained with an anti-α-tubu­lin an­ti­body (red). Ar­row­heads in­di­cate the lo­ca­tion of cen­tro­somes. (C, D) En­doge­nous γ-tubu­lin de­tec­tion in im­mu­no­elec­tron mi­croscopy us­ing 385Ab. Im­ages show the plasma mem­brane [PM], cy­tosol, nu­clear en­ve­lope [NE] and γ-tubu­lin bridges [γTB] of an in­ter­phase U2OS cell. (A-D) White and black boxes show the mag­ni­fied ar­eas dis­played in the in­sets. (E) The γ-string mesh­work is com­posed of treads (γ-string), which are at­tached to the plasma mem­brane [PM], found in the cy­toso­lic [C] com­part­ment and en­ter into the nu­clear com­part­ment [N] through the nu­clear en­ve­lope [NE]. In the nu­clear com­part­ment, γ-strings are in­ter­twined with the lam­ina mesh­work and with chro­matin. γ-Tubu­lin bridges [γTB] con­nect the cy­toso­lic and the nu­clear γ-tubu­lin pools cre­at­ing a nu­clear bound­ary of γ-tubu­lin [γTuB].
alt-text: Fig. 18
­vis egg ex­tracts demon­strate that lam­ina for­ma­tion oc­curs only around chro­matin (Lopez-Soler et al., 2001). A pos­si­ble ex­pla­na­tion for the dis­crep­ancy is that in the X. lae­vis egg ex­tracts, a cy­toso­lic γ-string bound­ary around the emerg­ing nu­cleus is first formed once the nu­cleus is as­sem­bled. Thus, the lack of a pre­formed cy­toso­lic γ-string bound­ary will pre­vent for­ma­tion of chro­matin-empty nu­clear like struc­tures. By con­trast, in cells, the cy­toso­lic γ-string bound­ary around chro­mo­somes and chro­matin-as­so­ci­ated γ-strings are never dis­as­sem­bled dur­ing cell di­vi­sion and thereby the bound­ary be­tween cy­toso­lic and chro­matin-as­so­ci­ated (-strings as­sures the for­ma­tion of a lam­ina around mi­totic chro­mo­somes.
Al­though the func­tions of γ-tubu­lin have been ex­ten­sively stud­ied over the past decades, there are not pre­vi­ous re­ports on γ-strings. There are two pos­si­ble rea­sons for this. First, most of stud­ies per­formed on γ-tubu­lin fo­cus on ex­am­in­ing the func­tion of γ-tubu­lin as a cen­tro­some and mi­cro­tubule or­ga­nizer and most of the an­ti­bod­ies com­mer­cially avail­able are se­lected for the recog­ni­tion of those γ-tubu­lin pools. Sec­ond, γ-tubu­lin is an abun­dant cel­lu­lar pro­tein that is found in both the cy­tosol and the nu­clear com­part­ment. This type of mesh­work is dif­fi­cult to stain and the stain­ing is dif­fi­cult to in­ter­pret as the an­ti­body stains both the cy­to­plasm and the nu­clear com­part­ment. An ex­am­ple of a mesh­work easy to stain is mi­cro­tubules. α- and β-tubu­lins form dis­tinct cy­toso­lic fibers (25 nm in di­am­e­ter), which are eas­ily de­tected with an­ti­bod­ies. The low con­cen­tra­tion of
sol­u­ble α- and β-tubu­lins sur­round­ing the ar­rays en­hances their stain­ing. In con­trast, γ-strings dif­fer from pre­vi­ous de­scribed mesh­work, as these are fine struc­tures that are dis­trib­uted in both the cy­toso­lic and the nu­clear com­part­ment. The lack of ar­eas with low con­cen­tra­tion of γ-strings makes it dif­fi­cult to de­tect their struc­ture.
The mech­a­nisms that reg­u­late γ-string for­ma­tion re­quire fur­ther analy­sis. Nonethe­less, in vitro γ-tubu­lin forms strings in the ab­sence of GTP and in cells, in the ab­sence of the GT­Pase do­main of γ-tubu­lin, the C-ter­mi­nal re­gion forms tubule-like struc­tures in a Ser 385-γ-tubu­lin de­pen­dent man­ner (Eklund et al., 2014). Thus, one can spec­u­late that γ-tubu­lin poly­mer­iza­tion de­pends on C-ter­mi­nal-to-C-ter­mi­nal bind­ing and the N-ter­mi­nal re­gion reg­u­lates its dy­nam­ics.
Our re­sults sug­gest that the in­ter­play be­tween the cy­toso­lic and the chro­matin-bound γ-strings across the nu­clear en­ve­lope plays an im­por­tant role dur­ing cell cy­cle. We pro­pose the fol­low­ing model. Dur­ing in­ter­phase, γ-tubu­lin bridges con­nect the cy­toso­lic and the nu­clear γ-tubu­lin pools. At the on­set of mi­to­sis, the lam­ina mesh­work is dis­rupted (Collas, 1999; Peter et al., 1990), but the γ-tubu­lin bound­ary around the mi­totic chro­mo­somes is main­tained. Dur­ing mi­to­sis, chro­matin-as­so­ci­ated γ-strings link the sis­ter chro­matids to the cy­toso­lic γ-string pool. Fi­nally, at anaphase/​telophase, the γ-tubu­lin bound­ary com­posed of cy­toso­lic and chro­matin as­so­ci­ated γ-strings forms a sup­port­ing scaf­fold that as­sist the for­ma­tion of the nu­clear en­ve­lope.
UNCORRECTED PROOF
Fig. 19. The C-ter­mi­nal re­gion of γ-tubu­lin as­sures the for­ma­tion of chro­matin-con­tain­ing nu­clei. (A) To­tal lysate from U2OS cells ex­press­ing C-γtubGFP334452 (CγTub), N-γtubGFP1333 (Nterm), GFP-γ-tubu­lin (γTub) or empty vec­tor (Cont.) were sep­a­rately im­muno­pre­cip­i­tated with anti-GFP or anti-lamin B. The ex­pres­sion lev­els of the var­i­ous re­com­bi­nant pro­teins were first an­a­lyzed by WB with an anti-GFP and reprobed with a mix­ture of two anti-γ-tubu­lin (T3320 and T6557) an­ti­bod­ies. The cel­lu­lose mem­branes con­tain­ing im­muno­pre­cip­i­tated GFP-tagged pro­teins were first an­a­lyzed by WB with an anti-lamin B an­ti­body and reprobed with an anti-lamin A/​C and GFP. Ar­row­heads in­di­cate the im­muno­pre­cip­i­tated GFP-fused pro­teins (n = 3). Graph shows the pro­tein con­cen­tra­tion of lamin B found in C-γtubGFP334452 and N-γtubGFP1333 im­muno­pre­cip­i­tates rel­a­tive to the lamin B con­cen­tra­tion found in GFP-γ-tubu­lin im­muno­pre­cip­i­tates ex­pressed in ar­bi­trary units (AU; mean ± s.d; n = 3, * p < 0.05). To ad­just for dif­fer­ences in pro­tein load­ing, the pro­tein con­cen­tra­tion of lamin B was de­ter­mined by its ra­tio with the im­muno­pre­cip­i­tated GFP-tagged pro­tein for each sam­ple. The pro­tein ra­tio in con­trol ex­tracts was set to 1. (B) DIC/​flu­o­res­cence im­ages of time-lapse from a U2OS cell that is sta­bly ex­press­ing both γTUBU­LIN shRNA and sh-re­sis­tant Cγ-tubGFP334452 (CγTub, green), and tran­siently ex­press­ing mCherry-lamin B1 (lamB1; red) with Hoechst 33258 stained chro­matin (blue), as in­di­cated. The im­age se­ries show cho­sen frames of the lo­ca­tion of Cγ-tubGFP334452 and lamin B1 dur­ing nu­clear as­sem­bly in a mi­totic cell. Im­ages were col­lected every 30 sec. Scale bars, 10 μm. See also movie S4.
alt-text: Fig. 19
Re­cently, an as­so­ci­a­tion of γ-tubu­lin with the nu­cle­o­porin MEL-28/​ELYS and the nu­clear en­ve­lope re-as­sem­bling GT­Pase, Ran, has been de­scribed (Yokoyama et al., 2014). In ad­di­tion, γ-tubu­lin as­so­ci­ated pro­teins such as αβ-tubu­lin and the γ-tubu­lin com­plex pro­tein 3-in­ter­act­ing pro­teins shape the nu­clear en­ve­lope (Batzenschlager et al., 2013; Xue et al., 2013). These data to­gether with the data pre­sented here sug­gest that the γ-tubu­lin mesh­work may be a struc­tur­ing scaf­fold that fa­vors the nu­cle­ation of var­i­ous pro­tein com­plexes nec­es­sary for nu­clear for­ma­tion.
Over­all, our re­sults sug­gest a novel mech­a­nism for how γ-tubu­lin co­or­di­nates cy­toso­lic and nu­clear events dur­ing cell cy­cle.
4. Materials and methods
4.1. cDNA, proteins and antibodies
Y. Zheng (Ma et al., 2009; Martin et al., 1998) and M. Klymkowsky (Dent et al., 1989) pro­vided lamin B3 pET-28a His6-tagged and anti-lamin B3, anti-Xgrip109 and anti-lamin II/​III an­ti­bod­ies (1:25). TUBU­LIN shRNA, sh-re­sis­tant γ-TUBU­LIN-1 gene, Nγ-tubGFP (γ-tubu­lin1333), Cγ-tubGFP (γ-tubu­lin334452), His6γtubu­lin, His6γ-Tub335451, His6E2F1 (Δ194-426) and GFP-SADBL were pre­pared as re­ported (Alvarado-Kristensson et al., 2009; Eklund et al., 2014; Hoog et al., 2011). Hu­man N-ter­mi­nal
GFP-tagged γtubu­lin and N-ter­mi­nal γ-tubu­lin frag­ment (His6γ-Tub1334) were ob­tained by PCR from γtubu­lin/​pEGFP-N1 (Alvarado-Kristensson et al., 2009) and sub­cloned in-frame into pEGFP-C1 (Clon­tech) or into pET21d (No­vagen), re­spec­tively, us­ing the fol­low­ing primers: 5GC­GAAGCTTC­GAT­GC­C­GAGGGAAAT­CATC3 and 5GCG­GAATTCT­CACT­GCTC­CTGGGT­GC­C­CCAGGA­GAT3 (γ-tubu­lin); 5GCG­GAATTCG­TAT­GC­C­GAGGGAAAT­CAT­CACC3 and 5CG­CAAGCTTGAC­CTGGGTGGGGT3 (His6γ-Tub1334). Hu­man RBM3 was am­pli­fied from hu­man cDNA and PCR RBM3 was sub­cloned in-frame into pEGFP-N1 (Clon­tech) us­ing the fol­low­ing primers: 5GCG­GC­TAGC­GAC­CAT­GTC­CTCT­GAA­GAAG­GAAAG3 and 5GC­GAAGCTTTTTCAT­GTTGT­CATAATTGTCTCT­GT3. The con­structs were ver­i­fied by se­quence analy­sis.
The fol­low­ing reagents were used: hu­man N-ter­mi­nal mCherry tagged Lamin B1/​pRe­ceiver-M55 (GeneCopoeia, Rockville, USA), anti-lamin B2, anti-lam­inA/​C, anti-cen­trin 2 (1:250; sc-27793), anti-hi­s­tone 2B (1:500), anti-hi­s­tone 3 (1:200), anti-GFP (sc-53882 and sc-8334; 1:500), mouse anti-γ-tubu­lin (1:400; sc-51715) and rab­bit anti-α-tubu­lin (1:400, all from Santa Cruz Biotech­nol­ogy, Dal­las, Texas, USA), mouse and rab­bit anti-γ-tubu­lin (1:250400, Sigma-Aldrich, Mu­nich, Ger­many; T5192, T3320 and T6557), anti-γ-tubu­lin (1:400; ab27074) and anti-His6 (1:1000, Ab­cam, Cam­bridge, UK), anti-α-tubu­lin (1:400, Mil­li­pore, Temec­ula, Cal­i­for­nia, USA), anti-
UNCORRECTED PROOF
Fig. 20. The N-ter­mi­nal re­gion of γ-tubu­lin leads to the for­ma­tion of chro­matin-empty nu­clei. (A) DIC/​flu­o­res­cence im­ages of time-lapse from a U2OS cell that is sta­bly ex­press­ing both γTUBU­LIN shRNA and sh-re­sis­tant N-γtubGFP1333 (NTerm, green), and tran­siently ex­press­ing mCherry-lamin B1 (lamB1; red) with Hoechst 33258 stained chro­matin (blue), as in­di­cated. The im­age se­ries show the lo­ca­tion of N-γtubGFP1333 and lamin B1 in a mi­totic cell that dur­ing nu­clear as­sem­bly tran­siently formed two nu­clear-like struc­tures, which lack chro­matin (ar­row­head and ar­row). Chro­matin-lack­ing nu­clear-like struc­tures (white bor­ders) are shown on the right im­ages. Graph shows the per­cent­age of filmed cells that formed chro­matin lack­ing nu­clear like struc­tures. Im­ages were col­lected every 30 sec. See also movie S5. (B) Fixed U2OS cells that are sta­bly ex­press­ing γTUBU­LIN shRNA and N-γtubGFP1333 were im­muno­flu­o­res­cence stained with an anti-lamin B an­ti­body (lam­inB; red) and nu­clei were de­tected with DAPI (blue). (A, B) Scale bars, 10 μm.
alt-text: Fig. 20
nu­clear pore com­plex (1:250, Co­v­ance, New Yersey, USA), anti-GFP (1:500) and anti-lamin B (1:500, both from Santa Cruz Biotech­nol­ogy, Dal­las, USA). Other reagents used in­cluded pro­tein G PLUS-sepharose (GE Health­care, Cleve­land, USA) and SDS-PAGE reagents (Bio-Rad, Cal­i­for­nia, USA). All other reagents were ob­tained from Sigma-Aldrich.
A rab­bit anti-γ-tubu­lin an­ti­body was gen­er­ated us­ing the phos­pho­pep­tide RVS­GLM­MAN­HTSISSLFE (phos­pho­ry­lated S un­der­lined; Pa­cific Im­munol­ogy, Cal­i­for­nia, USA). The anti-to­tal-γ-tubu­lin an­ti­body (1:400) was pu­ri­fied us­ing a ma­trix cou­pled co­va­lently to the non-phos­pho­ry­lated pep­tide, as de­scribed (Alvarado-Kristensson et al., 2009).
4.2. Manipulation of Xenopus laevis eggs and sperm
The Ethics Com­mit­tee of Lund Uni­ver­sity ap­proved the study (ref­er­ence num­ber: M 15111). In­ter­phase (CSF-ar­rested) egg ex­tracts were pre­pared in CSF-XB buffer (10 mM HEPES-KOH pH 7.7, 50 mM su­crose, 0.1 M KCl, 1 mM CaCl2, 2 mM MgCl2) sup­ple­mented with 0.1 mg/​ml cy­tocha­lasin B as de­scribed (Murray, 1991) with the fol­low­ing mod­i­fi­ca­tions to im­prove the qual­ity of the egg ex­tracts: ovu­la­tion was in­duced in­ject­ing 600 units of hu­man chori­onic go­nadotrophin (Sigma); paraf­fin oil was used; and, af­ter crush­ing, the su­per­natant was cen­trifuged at 10,000 g for 12 min to re­move re­main­ing de­bris. Fi­nally, in­ter­phase egg ex­tracts were ob­tained by sup­ple­men­ta­tion with the cal­cium ionophore A23187 (sigma) (Losada et al., 1998).
De­mem­branated sperm were pre­pared as pre­vi­ously de­scribed (Murray, 1991) us­ing ben­zo­caine 0.05% (W/​V) as an anes­thetic with the fol­low­ing mod­i­fi­ca­tions. To re­move pos­si­ble mem­brane frag
­ments; cy­toso­lic com­po­nents and con­t­a­m­i­nat­ing mi­cro­tubules, sperms were pel­leted by cen­trifu­ga­tion (18,000 g, 2 min) onto a 200 μl cush­ion con­tain­ing SuNaSp, 1.3 M and glyc­erol (Felix et al., 1994). Oc­ca­sion­ally, pre­vi­ous cen­trifu­ga­tion onto the glyc­erol cush­ion, sperm were pre­pared in the pres­ence of 5 μg/​ml col­cemid and 2.5 μg/​ml cy­tocha­lasin B (Murray, 1991). Re­ac­tions with in­ter­phase egg ex­tracts and sperm were in­cu­bated at 22 °C. Im­mun­ode­ple­tion were per­formed twice as de­scribed (Ma et al., 2009) us­ing each time 5 μl of anti-γ-tubu­lin T3320, anti-lamin B3 or no an­ti­body (con­trol de­ple­tion). De­mem­branated sperm was de­pleted from pro­teins (Ksperm) by treat­ment with 0.25 mg/​ml pro­teinase K for 17 min at 37 °C. Re­ac­tions were stopped by plac­ing sam­ples on ice and adding 30 mg/​ml glycine, 0.4 mM phenyl­methane­sul­fonyl flu­o­ride and 0.38% BSA.
For add-back ex­per­i­ments, hu­man His6-tagged fu­sion pro­teins (His6γ-tubu­lin, His6γ-Tub1334, His6γ-Tub335451 and His6E2F1 [Δ194-426]) and Xeno­pus His6-tagged lamin B3 (Ma et al., 2009) were ex­pressed as de­scribed (Hoog et al., 2011; Tsai et al., 2006). 100 ng of re­com­bi­nant pro­teins were added to the im­mun­ode­pleted egg ex­tracts or pre-in­cu­bated with Ksperm for 10 min prior to ad­di­tion of egg ex­tract. 500 sperm/μl ex­tract was used in as­sem­bly re­ac­tions. To an­a­lyze the func­tion­al­ity of the var­i­ous re­com­bi­nant His6γ-tubu­lin pro­teins, formed nu­clei were spun down (2,800 g) through a cush­ion of 30% su­crose (w/​v) in BAD (Mendez and Stillman, 2000) and pel­lets were lysed and an­a­lyzed as pre­vi­ously de­scribed (Alvarado-Kristensson et al., 2002).
To test the speci­ficity of the an­ti­bod­ies used, re­com­bi­nant His6γtubu­lin affin­ity pre-bound to Ni2+ affin­ity resin (Qi­a­gen) was in­cu­bated in the pres­ence of the an­ti­bod­ies T3320 or ab27074 for 2 h be­fore im­munos­tain­ing de­mem­branated sperm.
UNCORRECTED PROOF
4.3. Expression and purification of recombinant proteins
The hu­man C-ter­mi­nal His6-tagged fu­sion pro­teins (His6-γtubu­lin, His6-γ-Tub1334 and His6-γ-Tub335451) were ex­pressed in Es­cherichia coli BL21 (DE3) (Strat­a­gene) as de­scribed pre­vi­ously (Hoog et al., 2011). The Xeno­pus N-ter­mi­nal His6-tagged lamin B3 fu­sion pro­teins were pro­duced in E. coli BL21 (DE3) as de­scribed else­where (Tsai et al., 2006). In brief, ex­po­nen­tially grow­ing bac­te­ria bear­ing the plas­mid were main­tained at room tem­per­a­ture overnight (His6-tagged lamin B3) or at 37 °C for 1 h (His6-γtubu­lin, His6-γ-Tub1334 and His6-γ-Tub335451). Re­com­bi­nant pro­teins were pu­ri­fied un­der na­tive con­di­tions us­ing Ni2+ affin­ity resin (Qi­a­gen), in bind­ing buffer (His6-tagged lamin B3: 50 mM Tris-HCl pH 8.0, 25% su­crose, 1% Tri­tonX-100, 1 mM PMSF, and 5 mM im­i­da­zole; His6-γtubu­lin, His6-γ-Tub1334 and His6-γ-Tub335451: 20 mM Tris-HCl pH 7.9, 500 mM NaCl, 1 mM MgCl2, 0.25 μM GTP, 5 mM beta-mer­cap­toethanol (βME), 1 mM PMSF and 5 mM im­i­da­zole). Then the pro­teins were washed in 50 mM Tris-HCl pH 8.0 and 60 mM im­i­da­zole (His6-tagged lamin B3) or in 20 mM Tris-HCl pH 7.9, 1 mM MgCl2, 0.25 μM GTP, 5 mM βME and 60 mM im­i­da­zole (His6-γtubu­lin, His6-γ-Tub1334 and His6-γ-Tub335451) and eluted in 50 mM Tris-HCl pH 8.0 and 1 M im­i­da­zole (His6-tagged lamin B3) or in 20 mM Tris-HCl pH 7.9, 500 mM NaCl, 1 mM MgCl2, 0.25 μM GTP, 5 mM βME and 1 M im­i­da­zole (His6-γtubu­lin, His6-γ-Tub1334 and His6-γ-Tub335451). The pu­ri­fied pro­teins were ex­changed into XB (His6-tagged lamin B3: 10 mM HEPES, pH 7.7, 50 mM su­crose, 100 mM KCl, 0.1 mM CaCl2, and 5 mM EGTA) or TAB (His6-γtubu­lin, His6-γ-Tub1334 and His6-γ-Tub335451: 40 mM K-HEPES, 150 mM NaCl, 1 mM MgCl2, 1 mM EGTA and 1 mM DTT, at pH 7.2) buffer us­ing a dial­y­sis mem­brane (Spec­trum­Labs) and con­cen­trated in an Am­i­con Ul­tra Cen­trifu­gal Fil­ter (Mil­li­pore).
4.4. In vitro γ-string and lamin B3 polymerization assays
γ-Strings, lamin B3 fibers or both were poly­mer­ized by in­cu­bat­ing 20 μM His6-γ-tubu­lin (Alvarado-Kristensson et al., 2009; Eklund et al., 2014; Hoog et al., 2011; Ma et al., 2009), 600 nM His6-lamin B3 (Ma et al., 2009) or both pro­teins for 1 h at 22 °C with the fol­low­ing buffers: γ-tubu­lin as­sem­bly buffer (TAB), lamin B3 as­sem­bly buffer (LAB: TAB, sup­ple­mented with 10 mM su­crose, 20 mM KCl and 20 μM CaCl2). The abil­ity of His6-γ-tubu­lin to as­sist the for­ma­tion of His6-lamin B3 fibers was tested in LAB buffer. The as­sem­bled mesh­work was fixed us­ing 1% of glu­taralde­hyde for 5 min. In some ex­per­i­ments TAB was sup­ple­mented with 1 mM GTP.
4.5. Cell culture, transfection, immunoprecipitation and fractionation
NI­H3T3 mouse fi­brob­last and U2OS hu­man os­teosar­coma cells were cul­tured trans­fected and frac­tion­ated as re­ported (Alvarado-Kristensson et al., 2009; Eklund et al., 2014; Hoog et al., 2011). Pro­teins in pu­ri­fied cel­lu­lar frac­tions or cell lysates were im­muno­pre­cip­i­tated as de­scribed (Eklund et al., 2014) with the fol­low­ing mod­i­fi­ca­tions. Be­fore im­muno­pre­cip­i­ta­tion, chro­matin-as­so­ci­ated com­plexes were re­leased from chro­matin by re­sus­pend­ing the chro­matin pel­lets in chro­matin de­grad­ing buffer (2 units/μl ben­zonase [Sigma], 20 ng/μl Dnase I [Sigma], 50 mM Tris pH 7.5, 150 mM NaCl, 2 mM MgCl2) for 15 min, 22 °C. West­ern blot­ting was per­formed as re­ported (Alvarado-Kristensson et al., 2002).
Sta­bly trans­fected γTUBU­LIN shRNA U2OS cells and the var­i­ous U2OS cell lines sta­bly ex­press­ing GFP-γtubu­lin (C- or N-tagged γtubu­lin), NGFP-γ-tubu­lin or CGFP-γ-tubu­lin were ob­tained as de­scribed (Lindstrom et al., 2015).
To iso­late nu­clei, cells were pre-in­cu­bated for 20 min with cul­ture medium con­tain­ing 100 ng/​ml col­cemid and 5 μg/​ml cy­tocha­lasin B (37 °C, 5% CO2) to re­lease cy­toskele­tal com­po­nents from nu­clei and cy­to­plasm. Nu­clei of har­vested cells were in­cu­bated for 5 min in BAD (Mendez and Stillman, 2000) con­tain­ing 0.1% tri­ton X100. Nu­clei were fixed for 15 min at room tem­per­a­ture with 4 vol­umes of 10% formalde­hyde in BAD and then spun (2,400 g) onto cov­er­slips through a cush­ion of 30% su­crose (w/​v) in BAD.
4.6. Fluorescence imaging microscopy
Im­munos­tain­ing of de­mem­branated sperm were per­formed by first air-dry­ing sperm and there­after were fixed for 3 min with methanol/​ace­tone (1:1; v/​v) at 80 °C. Nu­clear as­sem­bly was ended by ad­di­tion of 4 vol­umes of 10% formalde­hyde in MMR (Murray, 1991) or 2% formalde­hyde sup­ple­mented with 0.25% Tri­ton-X100 in XBE2 (Losada et al., 1998). Fixed nu­clei were spun onto cov­er­slips through a cush­ion of 30% su­crose (Losada et al., 1998). Sub­se­quent im­munos­tain­ing was per­formed as de­scribed (Alvarado-Kristensson et al., 2009; Brown et al., 1995). The in­tegrity of the nu­clei formed was vi­su­al­ized by ex­clu­sion of 0.1 mg/​ml 155-kD tetram­ethyl­rho­damine isoth­io­cyanate (TRITC)-la­beled dex­tran, as de­scribed (O'Brien and Wiese, 2006).
Cells were cul­tured on cov­er­slips and to con­firm the pres­ence of γ-strings un­der var­i­ous fix­a­tion con­di­tions, we used three dif­fer­ent fix­a­tion pro­ce­dures. First, cells were fixed and per­me­abi­lized in methanol/​ace­tone (1:1; 80 °C, 5 min), sec­ond, cells were fixed in 4% paraformalde­hyde (PFA; RT, 5 min) and per­me­abi­lized in 0.1% Tri­ton-X100, and third, cells were fixed in 4% PFA-2% su­crose (RT, 3 min), fol­lowed by per­me­abi­liza­tion in methanol/​ace­tone (1:1; 80 °C, 3 min). Cells were in­cu­bated (1 h) with Alex­a480- or Cy3-la­belled sec­ondary an­ti­body (Jack­son). Cap­tur­ing of flu­o­res­cence and con­fo­cal im­ages were per­formed us­ing an Olym­pus IX73 mi­cro­scope (Olym­pus, Tokyo, Japan) and a Zeiss Axio Ob­server mi­cro­scope (Zeiss, Jena, Ger­many), re­spec­tively, as pre­vi­ously de­scribed (Alvarado-Kristensson et al., 2009; Hoog et al., 2011). Su­per-res­o­lu­tion im­ages were cap­tured with an ELYRA PS. 1 SIM/​PALM su­per-res­o­lu­tion struc­tured il­lu­mi­na­tion (SR-SIM; Zeiss). A min­i­mum of 100 nu­clei was counted in each sam­ple, and the per­cent­age of iso­lated nu­clei with lamin B1 stain­ing or of formed nu­clei was cal­cu­lated.
Near si­mul­ta­ne­ous GFP/​mCherry/​pm­Turquoise2/​DIC imag­ing se­quences were col­lected as de­scribed (Eklund et al., 2014; Lindstrom et al., 2015). Time-lapse im­ages were cap­tured every 2 min or 30 sec. Time in­ter­vals of the mi­totic processes were de­ter­mined by count­ing film frames. DNA was stained for 1 h with 1 μg/​ml Hoechst 33258. Quan­tifi­ca­tion of flu­o­res­cence of chro­matin-as­so­ci­ated GFP-tagged γ-tubu­lin con­structs was per­formed with Im­ageJ (Fiji) soft­ware (Na­tional In­sti­tutes of Health, USA).
4.7. Electron microscopy
For high pres­sure freez­ing, U2OS cells were seeded to 100% con­flu­ence onto car­bon coated (10 nm) 6 mm sap­phire discs (Le­ica) in 12-well dishes (Nunc). Cells were cryo-pre­served with high pres­sure freez­ing (HP­M100, Le­ica) fol­lowed by freeze sub­sti­tu­tion (Le­ica AFS2, Le­ica) for 48 h at 90 de­grees in Ace­tone with 0.1% Uranyl
UNCORRECTED PROOF
Ac­etate and em­bed­ded in Lowicryl with poly­mer­iza­tion at 25 de­grees for 48 h. 60 nm sec­tions were cut with Le­ica Ul­tra­cut UC7 (Le­ica, Vi­enna, Aus­tria) and col­lected on one whole for­m­var coated car­bon grids and 200 mesh Nickel grids. The sec­tions were pre-in­cu­bated for 30 min with pre-in­cu­ba­tion buffer (50 mM gly­sine, 0.1% sodium borhy­dride NaBH4, 0.05 M Tris pH 7.4, 0.1% Tri­ton), be­fore in­cu­ba­tion with poly­clonal anti-γ-tubu­lin or anti-α-tubu­lin an­ti­body in TBST (0,05 M Tris pH 7.4, 0.1% Tri­ton, 1% BSA) for 2 hours at room tem­per­a­ture, fol­lowed by 1 h in­cu­ba­tion with gold con­ju­gated pro­tein A (1:100, 10 nm gold; Agar Sci­en­tific, Es­sex, UK) in TBST. Fi­nal stain­ing was per­formed with fil­tered 0.5% uranyl ac­etate for 10 min.
For paraformalde­hyde fix­a­tion, cells grown on tis­sue cul­ture treated poly­car­bon­ate mem­branes (Corn­ing-Costar, NY, USA) were fixed with 4% paraformalde­hyde (5 min). Sam­ples were low tem­per­a­ture em­bed­ded in Lowicryl HM20 (EMS, PA, USA) in Le­ica AFS and sec­tioned into 50 nm ul­tra­thin sec­tions for trans­mis­sion elec­tron mi­croscopy on a Le­ica EM UC7 (Le­ica Mi­crosys­tems GmbH, Wet­zlar, Ger­many) be­fore trans­fer­ring to gold grids (EMS, PA, USA). Sec­tions were blocked in PBS and 0.5% BSA (30 min), in­cu­bated with poly­clonal anti-γ-tubu­lin or anti-γ-tubu­lin 385 an­ti­bod­ies (both 1:100, 1 h) fol­lowed by goat-anti-rab­bit IgG (20 or 10 nm gold; 1:20, 60 min; Agar Sci­en­tific, Es­sex, UK). Fi­nal stain­ing was per­formed with fil­tered 4% uranyl ac­etate. For mesh­work stain­ing, car­bon-coated cop­per grids were glow dis­charged. The poly­mer­iza­tion as­says (5 μl) were ap­plied on each grid for 5 min. Grids were twice washed with wa­ter for 1 min, and then 5 μl of freshly fil­tered 1% uranyl ac­etate were ap­plied for 30 s. Ex­cess stain was wicked off with fil­ter pa­per, and grids were al­lowed to dry at room tem­per­a­ture. Im­ages were ob­tained us­ing a Fei Tec­nai Spirit trans­mis­sion elec­tron mi­cro­scope (Fei, Hills­boro, Ore­gon, USA) and SIS Veleta (2 × 2k) CCD cam­era (Olym­pus, Tokyo, Japan).
4.8. Statistical analysis
All data are ex­pressed as means ± s.d., and sta­tis­ti­cal sig­nif­i­cance of the dif­fer­ences be­tween two groups or sev­eral groups was an­a­lyzed by paired Stu­dents t test: * p < 0.05, ** p < 0.01.
Declarations
Author contribution statement
Catalina Ana Rosselló, Lisa Lind­ström, Jo­han Glin­dre, Greta Ek­lund: Per­formed the ex­per­i­ments; An­a­lyzed and in­ter­preted the data.
Maria Al­varado-Kris­tens­son: Con­ceived and de­signed the ex­per­i­ments; Per­formed the ex­per­i­ments; An­a­lyzed and in­ter­preted the data; Wrote the pa­per.
Funding statement
This work was sup­ported by the Swedish Can­cer So­ci­ety; the Swedish child­hood can­cer foun­da­tion; the Royal Phys­io­graphic So­ci­ety in Lund; Gun­nar Nils­son; Crafo­ord­ska and the Skane Uni­ver­sity Hos­pi­tal in Malmö Can­cer Re­search Fund.
Competing interest statement
The au­thors de­clare no con­flict of in­ter­est.
Additional information
Sup­ple­men­tary con­tent re­lated to this ar­ti­cle has been pub­lished on­line at 10.​1016/​j.​heliyon.​2016.​e00166.
ACKNOWLEDGEMENTS
We thank Y. Zheng and M. Klymkowsky for reagents, Lund Uni­ver­sity Bioimag­ing Cen­ter and Core Fa­cil­ity for In­te­grated Mi­croscopy, Fac­ulty of Health and Med­ical Sci­ences, Uni­ver­sity of Copen­hagen for sup­port with elec­tron mi­cro­scope and Cen­tre for cel­lu­lar imag­ing at the Sahlgren­ska Acad­emy, Uni­ver­sity of Gothen­burg for sup­port with 3D su­per-res­o­lu­tion struc­tured il­lu­mi­na­tion mi­cro­scope and El­e­vate Sci­en­tific for ed­i­to­r­ial as­sis­tance.
References
UNCORRECTED PROOF
Please note that this text is not editable
Journal Id: YEXCR
Article Id: 166
e00166
Title:
Gamma-tubulin coordinates nuclear envelope assembly around chromatin
Authors: 
Catalina Ana Rosselló
Lisa Lindström
Johan Glindre
Greta Eklund
Maria Alvarado-Kristensson
S2405-8440(16)30323-1
Received Date: 22 April 2016
Revised Date: 10 July 2016
Accepted Date: 16 September 2016