Systems biology approach to study the role of miRNA in promoter targeting during megakaryopoiesis

Itishri Sahu, Rucha Hebalkar, Sonika Kar, SreeVidya T S, Usha Gutti, Ravi Kumar Gutti

Abstract

The distinct process of megakaryopoiesis requires occurrence of endomitosis for polyploidization of the megakaryocytes. Although, Cyclins, CDKs and have been described to regulate endomitosis, the exact mechanism still remains an enigma. miRNA which were otherwise known as post transcriptional gene silencers are now emerging with various noncanonical functions including gene regulation at pre-transcriptional level by miRNA binding at promoter region. Out of the many processes they regulate, miRNA have been manifested to play a role in megakaryocyte differentiation. In this study an attempt has been made to identify miRNA that could regulate cell cycle genes (Cyclins and CDKs) by targeting their promoters, during megakaryopoiesis. A new computational algorithm was implemented using Perl programming to identify putative targets of miRNA in CDK and Cyclin promoters. Perl script was also used to check nuclear localizing miRNA based on the presence of a consensus sequence. Real-time PCR was performed to analyze the expression of miRNA and their predicted targets in Dami vs. PMA treated Dami cells. Putative targets of miRNAs with longest, high complementarity matches in CDK/Cyclin promoters were obtained. We identified two significant miRNA, miR-1273g-3p and miR-619-5p with longest seed sequence matches. We further identified three main targets (CDK10, CDK11, Cyclin F) through which these two miRNA could regulate cell cycle during megakaryopoiesis. Our results reinforce the role of promoting targeting miRNA in regulation of cell cycle through certain CDK/Cyclins to support the process of endomitosis during megakaryopoiesis.

Introduction

Dependent kinase inhibitors (CDKs) have been implicated in the regulation of endomitosis [2– 4]. Studying regulation of cell cycle during megakaryopoiesis may provide a glimpse into how the endomitosis is facilitated. Evidently, a number of miRNAs are involved in regulating megakaryocyte differentiation. For instance, miR-125b positively regulates MK maturation by targeting cell cycle inhibitor p19INK4D at all stages of MK differentiation [5]. c-Myb, known as the negative regulator of megakaryocyte differentiation and CDK 4, CDK6 which regulate the G1/S phase transition are downregulated by miR-34a promoting the megakaryocyte differentiation [6]. Recent study suggests that miR-10a and 10b promote normal megakaryopoiesis by upregulating human platelet glycoprotein (GP Ibα) [7]. These 22-24nt long miRNAs are widely known to negatively regulate gene expression either by sequence specific mRNA degradation or translational repression. However, their non-canonical roles have started emerging. They are now reported to have activatory effect on gene expression at pre-transcriptional level by targeting the promoters. The first clue for the existence of such phenomenon came to light in 2006 in which synthetic dsRNA were found to activate gene expression by targeting promoter [8]. This phenomenon was termed as “RNA activation” (RNAa) and it was hypothesized that miRNA could also exhibit similar role. The earliest evidence was reported in 2008; hsa-miR-373 activated E-Cadherin and cold-shock domain-containing protein expression by targeting their promoter region with ~ 80% complementarity. [9] Subsequently, many other miRNAs having such activatory effect became evident. [10–12]
The miRNA-Ago complex is imported into the nucleus by Importin (IPO8), TNRC6A. The miRNA cause epigenetic changes at the promoter by recruiting chromatin modifying proteins leading to increase in H3K4 methylation or increase in H3K9/K27 methylation thereby altering the gene expression [13,14]. Two theories had been proposed for the mechanism through which this promoter targeting could happen: (i) miRNA directly bind to the chromosomal DNA like transcription factors and recruit various transcriptional regulators (ii) miRNA bind and target the non-coding transcripts overlapping the promoter region and mediate formation of complexes with proteins and chromosomal DNA to regulate the transcription. Schwartz JC et al., 2008 reported that dsRNA can either repress or activate gene expression and demonstrated that promoter-directed antigene RNAs (agRNAs) form a complex with AGO protein and bind to an antisense transcript overlapping the promoter. The antisense transcript–agRNA–AGO complex then acts as a scaffold for recruiting or redirecting other factors, such as heterogeneous nuclear ribonucleoprotein-k, RNA polymerase II and heterochromatin protein 1γ [15]. Follwing this, miR-423-5p was proved to cause transcriptional silencing of progesterone receptor gene by targeting RNA transcript overlapping its promoter [16].
Promoter targeting requires nuclear transport of miRNA and studies have indicated their localization in the nucleus [17,18]. These nuclear localizing miRNA may contain a specific signal sequence that guides them into the nucleus [19]. MiRNA have also been shown to interact with transcription factors and it is possible that miRNA direct transcription factors towards particular gene by directing it with the help of longer seed sequence complementarity [20,21]. Higher base complementarity adds to both complexity and specificity as TFs recognize rather short length of sequence [22].
Based on the above findings, we hypothesized the existence of role of miRNA in regulating megakaryopoiesis at pre-transcriptional level through promoters of cell cycle genes (Cyclins, CDKs) as targets. Probable targets/matches of all human mature miRNAs in cell cycle gene promoters were predicted using an algorithm using Perl programming. Referring to the previous reports of promoter targeting miRNA and their probable interaction with transcription factors, stringency of perfect and higher base complementarity (6-19 nt seed sequence match) was set while finding miRNA-promoter matches. Higher complementarity matches and frequency of matches in the Cyclin and CDK promoters were further analyzed. Probable nuclear localizing miRNAs (presence of ASUS motif at their 3’ end as stated by Hwang and colleagues [19] ) which also had long match (>11nt) with Cyclin/CDK promoter(s) were identified. Significant longest matching miRNAs and their Cyclin/CDK targets were further studied in the system of differentiated megakaryocytes. Materials and Methods

Dataset Collection

The promoter sequences corresponding to 1000bp upstream of Transcription Start Site of all human CDKs (25) and Cyclins (21) were extracted from Ensemble server (www.emsembl.org) and stored as text files [23]. A consolidated list (Table 1A and 1B) of the 25 Cyclins and 21 CDKs along with their gene IDs, gene location and location of the 5’flanking promoter sequence was prepared. Complete set of mature miRNA sequences were obtained from miRBase website (www.mirbase.org) in a FASTA file. Using a Perl script, only the human mature miRNAs (2578) were extracted and stored in a text file.

Base Composition

To understand any broad similarities between the sequences, theA/T/G/C contentin promoter sequences and A/U/G/C content in miRNA were calculated using a Perl script.

miRNA (seed sequence) – promoter matching

In congruence with the previous reports of promoter targeting miRNA and their probable interaction with transcription factors, stringency of perfect and higher base complementarity (619 nt seed sequence match) was set while finding miRNA-promoter matches. Prior to finding matches, complementary (forward) and reverse complementary (reverse) sequences of all Cyclin and CDK promoters were generated using Perl. Substrings of miRNA sequence of length 6nt to 19ntwere generated using “substr” function in Perl. Presence of substring was checked in each promoter (forward& reverse) sequence. The individual output files containing miRNA name, the matched sequence and frequency of matches were generated.The outputs obtained were pre processed using Microsoft Excel.

Nuclear localizations of miRNAs

All the miRNAs were checked for their probable nuclear localization by the presence of ASUS motif (as stated by Hwang and colleagues[19]) at their 3’ end using a Perl script.miRNA having long matches (>11nt ) with cell cycle gene promoters were checked for the presence of ASUS motif.

Cell Culture

Human megakaryoblastic cell line, Dami was cultured in RPMI-1640 medium supplemented with 10% Fetal Bovine Serum (FBS) and 1% antibiotics. The cells were grown in humidified incubator at 37oC in the presence of 5% carbon dioxide.

PMA treatment

Dami cells were cultured in above mentioned media which was also supplemented with Phorbol 12-myristate 13-acetate (100ng/mL), for 72hrs to establish a model for megakaryocyte differentiation [24]. The differentiation of Dami cells into megakaryocytes upon PMA treatment was confirmed by analyzing the increase in expression of megakaryocyte markers (CD61, CD41) by real time PCR as well as flow cytometry.

Real time PCR

RNA was isolated from Dami cells and 72hrs PMA treated Dami cells using miRNeasy mini kit (Qiagen) and the cDNA was synthesized using first- strand cDNA synthesis kit (Takara) according to the manufacturer’s instructions. This cDNA was used as a template for real time PCR which was carried out using SYBR Green FAST qPCR Master Mix (Kappa Biosystems) in an ABI step one plus detection system (Applied Biosystems). The thermal cycling parameters were set as: initial denaturation at 94oC for 10min, followed by 40 cycles of denaturation at 94oC for 15 sec, annealing for 30sec. and extension at 72oC for 30 sec. Specific primers used for quantification are listed in theSupplementary Table 1.The PCR product’s specificity was verified by melting curve and agarose gel analysis. The results presented are from three individual experiments, in which each sample was assayed in triplicate, normalized to the level of GAPDH (internal control).
For miRNA expression studies, miRNA was isolated using miRNeasy mini kit (Qiagen). Isolated miRNA was reverse transcribed using miScript II RT Kit (Qiagen). Quantification of mature miRNAs (miR-1273g-3p and miR-619-5p) was done using miScript primer assay kit (Qiagen) and ABI Step One plus detection system. Values were normalized against U6 (miRNA internal control).

Statistical analysis

Two-tailed Student’s t-test was used to determine the statistical significance in in-vitro studies. Data was expressed as mean ± SD. P-value<0.05was considered to be statistically significant.

Results and Discussion

Base Analysis

Significant information can be obtained by comparing the base composition of promoters and miRNAs. Thus, base composition for miRNAs and all thepromoters were calculated. The graphs were plotted against percentage (%) of A/T(U)/G/C and promoters or miRNAs. Figure 1A, 1B, and 1C were used to interpret the uniformity or differences in the base composition of all the data. The average AT/GC ratio for the Cyclins was calculated as 0.944. A few exceptions were Cyclin E1, Cyclin F, and Cyclin K whose AT/GC ratio was calculated as 0.42, 0.42, 0.29 and 0.50 respectively. They were found to have a higher CG content than AT content. Similarly, the average AT/GC ratio for CDKs was calculated as 0.92.A few exceptions are CDK 3, 9 and 12 whose AT/GC ratio was calculated as 0.56, 0.42 and 0.58 respectively. They had noticeably higher CG content than the AT content.Higher GC content in the promoters might be due to presence of CpG islands in these important regulatory regions. For 20 random miRNAs, GC content was found to be slightly higher than the AU content (Figure 2C). The miRNAs taken individually, varied in their AU/GC ratio. This may have been observed due to complete absence of a particular base in the miRNA sequence. For instance, it was seen that U was absent in miR3960, C was absent in let-7a-5p, A was absent in miR-1469.Interestingly, if all miRNAs were taken as a whole, AU/GC ratio obtained was 0.91(~ 1). Deviations from the average AT/GC ratio or AU/GC ratio may have occurred because of non-uniform distribution of bases in the individual sequences.

miRNA (seed sequences) - promoter matches

The 6nt-19nt matches between the miRNA and the promoters (complementary and reverse complementary) were found out by a Perl script. The frequencies of the exact matches were tabulated (seed-wise for complementary and reverse complementary strands respectively) as shown in Supplementary Table 2A & 2B.Frequencies of match for 5 random Cyclins and CDKs were plotted (Fig 2A & 2B). 10 nucleotide match was kept as a threshold while plotting as lower nucleotide frequencies of matches were very high (>8000 for Cyclins and >100 for CDKs) and difficult to represent graphically. Considering the longest match, 17 nucleotide match was kept as the upper limit while plotting. It was observed that the frequency of match decreased with increasing seed sequence length and the frequency of match in the reverse strand was higher than that in forward strand in majority of the cases. Phase wise miRNA targeting thepromoter sequence of cyclins and CDKs were identified (Table 2). Amongst these, miR-6858-5p was found to have a high 15 nucleotide match with CDK4 (required during G1) and miR-1469 had a 16 nucleotide match with Cyclin E1 (required during G1/S transition).
The highest seed sequence match between 11nt-17nt for all the promoters were recorded (Supplementary Table 3).The longest match amongst all CDKs and Cyclins was 17nt long which was between CDK10 (reverse) and miR-1273g-3p. A single miRNA i.e. miR-619-5p showed matches with highest number of promoters (i.e. 6 promoters at the reverse strand). The promoters with which the miR-619-5p matched (along with the cases of the longest length matched) were: CDK11A -15nt match, CDK11B-15nt match, CDK12-15nt match, CDK19-14nt match, Cyclin L1-16nt match and Cyclin Y-15nt match. RNAhybrid sever [25] was used to calculate the minimum free energy of binding of miR-1273g-3p and miR-619-5p to their predicted targets (Table 3). All the matches were found to be energetically favorable.

miRNA localization

In this study, the presence of ASUS motifs required for nuclear localization as reported by Hwang and colleagues(towards the last 10 bases of 3′ end) in whole (2578) miRNAdataset was checked (using a Perl script). 320 such miRNAs were found that had ASUS motifs towards the last 10 bases of 3′ end. In previous studies by Jeffries et al., among the 100most strongly detected miRNAs, a set of 21 miRNAs were found to have high concentration rank in the nucleus; where as another set of 31 miRNAs had the same in the cytoplasm [26]. When compared with the list of 21 miRNAs that targeted into nucleus, 8 of the 21 miRNAs were found in the list of the 320 miRNAs set obtained above that contained ASUS motifs at their 3′ end. Those were listed out to be miR-19a-3p, miR-197-5p, miR-30c, miR-30b-5p, miR-374a-5p, miR-193b-3p, miR-590-5p and miR-374b-5p. However, when compared with the 31 miRNAs that targeted the cytoplasm, none of the miRNAs were found in the ASUS containing (all 320) miRNA list which clearly suggests some role of ASUS in miRNA nuclear localization. Also, the miRNAs with ASUS, targeting CDKs & Cyclins were found out. A total of nine CDKs and seven Cyclins were noticed to be targeted by ASUS containing miRNAs as shown in Table 4.These need to be experimentally validated.

Expression of miRNA &their targets (and their significance)

The functions of longest matching miRNA (miR-1273g-3p) and miRNA matching highest number of promoters (miR-619-5p) are still not explored much. Although these were absent in the list of miRNAs with ASUS, targeting CDKs & Cyclins, they show a high potential as promoter regulatory elements considering their long, perfect complementarity and energetically favorable matches. Interestingly, miR-619-5p was found to be one of the 15 novel mature miRNAs derived from known miRNA precursors that had been detected in human platelets [27]. The endogenous expression of miR-1273g-3p and miR-619-5p was checked in Dami vs. 72hrs PMA treated Dami (differentiated MKs) cells using real time PCR. Both the miRNAs were found to be significantly downregulated (~10 fold) (Figure 3) Expression of targets of these two miRNAs with lower as well as higher complementarity was checked in Dami vs. PMA treated Dami cells. Significant differential expression(downregulation)was observed only in the case of few genes; CDK10, CKD11, Cyclin F.(Figure 4)The variable expression of these genes could be relevant in megakaryocyte system because: 1. Complex of CDK10/Cyclin M is expressed during G2/M phase and is responsible for degradation of ETS2 transcription factor [28]. This ETS2 transcription factor has been reported to promote megakaryocyte differentiation [29,30]. The expression of ETS2 was checked in Dami vs. PMA treated Dami cells and was observed to be upregulatedupon PMA treatment. 2. Reduced protein expression of CDK11 has been reported to lead to impaired cytokinesis [31]. Cytokinesis has to cease during endomitosis during megakaryocyte differentiation and maturation. 3. Cyclin F is a member of F-box protein family and in contrast to most Cyclins, it does not activate any CDK. siRNA-mediated depletion of Cyclin F leads to mitotic abnormalities like multipolar spindles which are observed during endomitosis [32]. It can be theorized that during megakaryocyte differentiation, downregulation of miR-1273g-3p leads to dowregulation of its potential promoter target, CDK10 which could further cause relative upregulation of ETS2 transcription factor and hence, promote megakaryocyte differentiation. Similarly, it can be speculated that decrease in the levels of miR1273g-3p and miR-619-5p could also possibly lead to downregulation of other potential targets such as CDK11 and Cyclin F, which aid in the process of endomitosis. Further experimental proof is necessary to reveal the localization of the two miRNA, their specificity to target the respective promoters and their mechanism of action.

Conclusion

We developed a different approach to unravel the regulation of megakaryopoiesis by miRNA at pre-transcriptional level. Perl script was developed to find the putative targets of human mature miRNA in Cyclin and CDK promoters. Analysis of the outputs of Perl algorithm revealed high complementarity miRNA which could target the cell cycle gene promoters. miRNA that could target in during cell cycle phase wise fashion were obtained. Interestingly, two miRNA; (i) miR1273g-3p having longest match with CDK10 and (ii) miR-619-5p matching 6 promoters with high complementarity (above 14nt) were thought to be significant results which required further experimental validation in megakaryocyte system. miR-619-5p was previously reported to be one of the novel miRNA expressed in platelets. Expression of these two miRNA was found to be downregulated (~10 fold) during MK differentiation. Amongst all the targets of these two miRNA, CDK10, CDK11 and Cyclin F displayed differential expression (downregulated) during MK differentiation. This becomes congruent with the fact that miRNA can activate gene expression.It can be suggested that CDK10 downregulation has a positive effect on ETS2 which facilitates megakaryocyte differentiation. CDK11 and Cyclin F might be vital for endomitosis to take place during the MK differentiation.However, experimental proof including time series miRNA profilingis required to confirm, respective Cyclin and CDK promoters as the miRNA targets during megakaryopoiesis.

References

[1] Tozawa K, Ono-Uruga Y, Matsubara Y. Megakaryopoiesis. Clin Exp Thromb Hemost 2014;1:54–8. doi:10.14345/ceth.14014.
[2] Sun S, Zimmet JM, Toselli P, Thompson A, Jackson CW, Ravid K. Overexpression of cyclin D1 moderately increases ploidy in megakaryocytes. Haematologica 2001;86:17–23.
[3] Eliades A, Papadantonakis N, Ravid K. New roles for cyclin E in megakaryocytic polyploidization. J Biol Chem 2010;285:18909–17. doi:10.1074/jbc.M110.102145.
[4] Gilles L, Guie R, Bluteau D, Lacout C, Debili N. P19 INK4D links endomitotic arrest and megakaryocyte maturation and is regulated by AML-1 2017;111:1–3. doi:10.1182/blood2007-09-113266.
[5] Qu M, Fang F, Zou X, Zeng Q, Fan Z, Chen L, et al. miR-125b modulates megakaryocyte maturation by targeting the cell-cycle inhibitor p19(INK4D). Cell Death Dis 2016;7:e2430. doi:10.1038/cddis.2016.288.
[6] Navarro F, Gutman D, Meire E, Cáceres M, Rigoutsos I, Bentwich Z, et al. miR-34a contributes to megakaryocytic differentiation of K562 cells independently of p53. Blood 2009;114:2181–92. doi:10.1182/blood-2009-02-205062.
[7] Zhang Z, Ran Y, Shaw T, Peng Y. MicroRNAs 10a and 10b Regulate the Expression of Human Platelet Glycoprotein Ibα for Normal Megakaryopoiesis. Int J Mol Sci 2016;17:1873. doi:10.3390/ijms17111873.
[8] Li L-C, Okino ST, Zhao H, Pookot D, Place RF, Urakami S, et al. Small dsRNAs induce transcriptional activation in human cells. Proc Natl Acad Sci U S A 2006;103:17337–42. doi:10.1073/pnas.0607015103.
[9] Place RF, Li L-C, Pookot D, Noonan EJ, Dahiya R. MicroRNA-373 induces expression of genes with complementary promoter sequences. Proc Natl Acad Sci U S A 2008;105:1608–13. doi:10.1073/pnas.0707594105.
[10] Majid S, Dar A, Saini S, Yamamura S. microRNA-205 directed transcriptional activation of tumor suppressor genes in prostate cancer. Cancer 2010;116:5637–49. doi:10.1002/cncr.25488.MicroRNA-205.
[11] Huang V, Place RF, Portnoy V, Wang J, Qi Z, Jia Z, et al. Upregulation of Cyclin B1 by miRNA and its implications in cancer. Nucleic Acids Res 2012;40:1695–707. doi:10.1093/nar/gkr934.
[12] Kang MMR, Park KKH, Yang J-OJ, Lee CCW, Oh SSJ, Yun J, et al. miR-6734 UpRegulates p21 Gene Expression and Induces Cell Cycle Arrest and Apoptosis in Colon Cancer Cells. PLoS One 2016;11:e0160961. doi:10.1371/journal.pone.0160961.
[13] Huang V, Li LC. miRNA goes nuclear. RNA Biol 2012;9:269–73. doi:10.4161/rna.9.3.19354.
[14] Catalanotto C, Cogoni C, Zardo G. MicroRNA in control of gene expression: An overview of nuclear functions. Int J Mol Sci 2016;17. doi:10.3390/ijms17101712.
[15] Schwartz JC, Younger ST, Nguyen N-B, Hardy DB, Monia BP, Corey DR, et al. Antisense transcripts are targets for activating small RNAs. Nat Struct Mol Biol 2008;15:842–8. doi:10.1038/nsmb.1444.
[16] Younger ST, Corey DR. Transcriptional gene silencing in mammalian cells by miRNA mimics that target gene promoters. Nucleic Acids Res 2011;39:5682–91. doi:10.1093/nar/gkr155.
[17] Liao JY, Ma LM, Guo YH, Zhang YC, Zhou H, Shao P, et al. Deep sequencing of human nuclear and cytoplasmic small RNAS reveals an unexpectedly complex subcellular distribution of mirnas and tRNA 3′ trailers. PLoS One 2010;5. doi:10.1371/journal.pone.0010563.
[18] Park CW, Zeng Y, Zhang X, Subramanian S, Steer CJ. Mature microRNAs identified in highly purified nuclei from HCT116 colon cancer cells. RNA Biol 2010;7:606–14. doi:10.4161/rna.7.5.13215.
[19] Hwang H-W, Wentzel EA, Mendell JT. A Hexanucleotide Element Directs MicroRNA Nuclear Import. Science (80- ) 2007;315:97–100. doi:10.1126/science.1136235.
[20] Korla K, Arrigo P, Mitra CK. Promoters, toll like receptors and microRNAs: A strange association. Indian J Biochem Biophys 2013;50:169–76.
[21] Dannemann M, Prüfer K, Lizano E, Nickel B, Burbano HA, Kelso J. Transcription factors are targeted by differentially expressed miRNAs in primates. Genome Biol. Evol., vol. 4, 2012, p. 552–64. doi:10.1093/gbe/evs033.
[22] Stewart AJ, Plotkin JB. Why transcription factor binding sites are ten nucleotides long. Genetics 2012;192:973–85. doi:10.1534/genetics.112.143370.
[23] Flicek P, Amode MR, Barrell D, Beal K, Billis K, Brent S, et al. Ensembl 2014. Nucleic Acids Res 2014;42. doi:10.1093/nar/gkt1196.
[24] Greenberg SM, Rosenthal DS, Greeley TA, Tantravahi R, Handin RI. Characterization of a BMS-265246 new megakaryocytic cell line: the Dami cell. Blood 1988;72:1968–77.
[25] REHMSMEIER M. Fast and effective prediction of microRNA/target duplexes. RNA 2004;10:1507–17. doi:10.1261/rna.5248604.
[26] Jeffries CD, Fried HM, Perkins DO. Nuclear and cytoplasmic localization of neural stem cell microRNAs. RNA 2011;17:675–86. doi:10.1261/rna.2006511.
[27] Plé H, Landry P, Benham A, Coarfa C, Gunaratne PH, Provost P. The Repertoire and Features of Human Platelet microRNAs. PLoS One 2012;7. doi:10.1371/journal.pone.0050746.
[28] Guen VJ, Gamble C, Flajolet M, Unger S, Thollet A, Ferandin Y, et al. CDK10/cyclin M is a protein kinase that controls ETS2 degradation and is deficient in STAR syndrome. Proc Natl Acad Sci U S A 2013;110:19525–30. doi:10.1073/pnas.1306814110.
[29] Rainis L, Toki T, Pimanda JE, Rosenthal E, Machol K, Strehl S, et al. The proto-oncogene ERG in megakaryoblastic leukemias. Cancer Res 2005;65:7596–602. doi:10.1158/00085472.CAN-05-0147.
[30] Ge Y, LaFiura KM, Dombkowski AA, Chen Q, Payton SG, Buck SA, et al. The role of the proto-oncogene ETS2 in acute megakaryocytic leukemia biology and therapy. Leukemia 2008;22:521–9. doi:10.1038/sj.leu.2405066.
[31] Wilker EW, van Vugt M a TM, Artim S a, Huang PH, Petersen CP, Reinhardt HC, et al. 14-3-3Sigma Controls Mitotic Translation To Facilitate Cytokinesis. Nature 2007;446:329–32. doi:10.1038/nature05584.
[32] D’Angiolella V, Donato V, Vijayakumar S, Saraf A, Florens L, Washburn MP, et al. SCF(Cyclin F) controls centrosome homeostasis and mitotic fidelity through CP110 degradation. Nature 2010;466:138–42. doi:10.1038/nature09140.