Maternal age effect on mouse oocytes: new biologicalinsight from proteomic analysis Caroline Schwarzer*, Marcin Siatkowski1,2,*, Martin J Pfeiffer, Nicole Baeumer3,Hannes C A Drexler4, Bingyuan Wang, Georg Fuellen1,2 and Michele Boiani Max Planck Institute for Molecular Biomedicine, Ro¨ntgenstraße 20, D-48149 Mu¨nster, Germany, 1DZNE, German Centre for Neurodegenerative Disorders, Gehlsheimer Straße 20, D-18147 Rostock, Germany, 2Institute for Biostatistics and Informatics in Medicine and Ageing Research, Rostock University Medical Center,Ernst Heydemann-Str. 8, D-18057 Rostock, Germany, 3Arrows Biomedical GmbH, Gievenbecker Weg 11,D-48149 Mu¨nster, Germany and 4Max Planck Institute for Molecular Biomedicine, Bioanalytical Mass SpectrometryFacility, Ro¨ntgenstraße 20, D-48149 Mu¨nster, Germany Correspondence should be addressed to M Boiani; Email: or to G Fuellen; Email: *(C Schwarzer and M Siatkowski contributed equally to this work) The long-standing view of ‘immortal germline vs mortal soma' poses a fundamental question in biology concerning how oocytes age inmolecular terms. A mainstream hypothesis is that maternal ageing of oocytes has its roots in gene transcription. Investigating the proteinsresulting from mRNA translation would reveal how far the levels of functionally available proteins correlate with mRNAs and would offernovel insights into the changes oocytes undergo during maternal ageing. Gene ontology (GO) semantic analysis revealed a high similarityof the detected proteome (2324 proteins) to the transcriptome (22 334 mRNAs), although not all proteins had a cognate mRNA.
Concerning their dynamics, fourfold changes of abundance were more frequent in the proteome (3%) than the transcriptome (0.05%),with no correlation. Whereas proteins associated with the nucleus (e.g. structural maintenance of chromosomes and spindle-assemblycheckpoints) were largely represented among those that change in oocytes during maternal ageing; proteins associated with oxidativestress/damage (e.g. superoxide dismutase) were infrequent. These quantitative alterations are either impoverishing or enriching.
Using GO analysis, these alterations do not relate in any simple way to the classic signature of ageing known from somatic tissues.
Given the lack of correlation, we conclude that proteome analysis of mouse oocytes may not be surrogated with transcriptome analysis.
Furthermore, we conclude that the classic features of ageing may not be transposed from somatic tissues to oocytes in a one-to-onefashion. Overall, there is more to the maternal ageing of oocytes than mere cellular deterioration exemplified by the notoriousincrease of meiotic aneuploidy.
Reproduction (2014) 148 55–72 (Thus, ageing of oocytes may The ‘maternal age effect' in reproduction, characterized be viewed as a life-long maintenance of cellular by a negative relationship between maternal age and homeostasis in the same cell, unlike ageing of the male reproductive success, is a poorly understood phenom- germline. Some molecular factor(s) within oocytes might enon. Several hypotheses have been proposed. While deteriorate as the potential mother ages, compromising, the receptivity of the uterus and the ovarian reserve of for example, the function of the chromosomal apparatus follicles can explain the ‘maternal age effect' in part, it is or the ability to scavenge reactive oxygen species known that age-related decline of a female's fertility is resulting from mitochondrial reactions ( also rooted in the quality and developmental potential of her oocytes. Unlike the male germline, the bulk of The synthesis of the aforementioned molecular oocytes do not have a gonial stem cell population.
factor(s) relies on gene transcription that oocytes largely Oocytes spend most of their time quiescent in primordial perform during follicular growth prior to ovulation, and follicles, mature over days or weeks during follicular the transcriptional activities of oocytes may be influ- growth and then become quiescent again near the enced by maternal age. Microarray and RT-PCR methods time of ovulation, when gene transcription is silenced have revealed that maternal ageing is accompanied q 2014 Society for Reproduction and Fertility ISSN 1470–1626 (paper) 1741–7899 (online) Online version via C Schwarzer, M Siatkowski and others by changes in the levels of oocytic mRNAs involved in and should harbour the majority of all oocyte proteins mitochondrial function, apoptosis, oxidative stress, cell (although with different relative abundances). Practically, cycle regulation, chromosome stability and epigenetic F9 cells allow for the efficient metabolic labelling of the modification, in both mouse and human oocytes SILAC reference in vitro, overcoming the difficulty of directly labelling oocytes in vivo.
We analysed maternal ageing of mouse oocytes on the protein level, using SILAC technology and high- However, searching these studies for culprits resolution MS, to define its signature at a level closer elicits a list of candidate genes that is quite short, to phenotype than mRNA. These oocytes had aged up to featuring Bcl2 and Bax (mitochondrial function and 1 year inside the ovaries and were analysed immediately apoptosis); Txndc9 (Apacd), Sod1 and Txn1 (oxidative after ovulation so as to appreciate the effect of maternal stress); Mad2l1 (Mad2) and Bub1 (spindle assembly age on the oocyte. This should not be confused with the checkpoint (SAC)); Atrx, Brca1, Numa1 and Smc1b post-ovulatory ageing of oocytes in the oviduct or in a (spindle assembly and chromosome integrity/stability); culture medium (We compared and Dmap1, Dnmt1, Dnmt3A and Hdac1/2 (epigenetic and confronted the quantitative protein data with the modification). Furthermore, the fold-change of these predictions of mRNA ageing studies of oocytes and mRNAs in oocytes during maternal ageing is low somatic cells. The analysis of two maternal age (sometimes as low as 1.4- to 1.5-fold).
transitions (puberty to mature age, first; mature age to The hypothesis that maternal ageing of oocytes has its climacterium, second) allowed us to disregard proteins roots in aberrant gene transcription, as opposed to other that keep steady throughout life and to focus on those steps of the gene expression cascade, is largely proteins whose abundance is changing, i.e. age unproven and is difficult to reconcile with the transcrip- regulated in the second age transition. Overall, corre- tional silencing that occurs in oocytes near the time of lation between quantitative changes in the proteome and ovulation. Advances in ‘omics' research show that in the transcriptome was nearly nil. Gene expression and intermediate steps of the gene expression cascade, gene ontology (GO) analyses revealed a distinction as well as post-translational protein modification and in the cellular components and biological processes degradation, can affect levels of functionally available (BPs) affected by ageing in oocytes compared to somatic proteins independent of transcription ( cells. Proteins associated with the nucleus were featured Thus, we consider that the protein level may offer predominantly among those that changed (declined) in novel insights into the changes in gene expression that oocytes during maternal ageing, thereby fulfilling a oocytes undergo during maternal ageing. The proteome, classic expectation of oocyte ageing, i.e. progressive loss an accessible ‘missing link' between transcriptome and of precision in chromosome maintenance and segre- phenotype, can be analysed using high-resolution mass gation. By contrast, the number of changing proteins spectrometry (MS)-based proteomics, which allows associated with oxidative stress/damage was very small.
for the accurate identification of thousands of proteins Our first conclusion is that proteome analysis of mouse in somatic tissues during in vivo ageing ( oocytes may not be surrogated with transcriptome analysis: the two molecular portraits do not correlate.
Previous MS studies on mice analysed the proteome of Our second conclusion is that the classic features of mature oocytes and ageing may not be transposed from somatic tissues processes of oocyte maturation ( to oocytes in a one-to-one manner. Overall, there is more Monti et al., 2013) in to the maternal ageing of oocytes than a mere cellular young donors, but not oocyte quality and composition deterioration exemplified by the notorious increase in during maternal ageing. As MS is not inherently quan- meiotic aneuploidy. This is in accord with our functional titative, relied on the simultaneous observations of the increased ability of old oocytes to comparison of signal intensities between replica samples support blastocyst formation, although reduced post- from young donors, in a so-called label-free approach that implantation development is predicted as a result of requires high numbers of oocytes. Scarce samples, such as oocytes from aged donors, are hardly amenable to thereplica collection of hundreds if not thousands of oocytes.
These scarce specimens can be quantified using a defined Materials and methods amount of isotopically labelled reference, which is added(‘spike-in') to the non-labelled oocyte lysate prior to Gene, mRNA and protein nomenclature processing for MS. This method is called SILAC (stable We followed the rules of the Jackson Laboratory published as of isotope labelling of amino acids in cell culture; ). F9 embryonal carcinoma (EC) cells are Briefly, gene names are written appropriate as a labelled reference for oocytes as they can in italic format (e.g. Hprt), mRNA names are written in italic easily be cultured feeder-free, have stem cell properties format and also specified to be mRNA (e.g. Hprt mRNA), Reproduction (2014) 148 55–72 SILAC proteome analysis of mouse oocyte ageing protein names are not written in italic format and only uppercase letters were used (e.g. HPRT).
For each age group, 20 oocytes were collected in biologicaltriplicates. Microarray data of the MII oocytes were obtained using the Agilent platform, as described )and deposited at This mouse ageing study was performed in accordance with the recommendations of the Federation of Laboratory Animal RNA was extracted using the ZR RNA Microprep Kit (Zymo Science Associations (FELASA) and with the ethical permit Research Corporation, Irvine, CA, USA). A two-round linear issued by the Landesamt fu¨r Natur, Umwelt und Verbrau- amplification protocol employing a linear two-step TargetAmp cherschutz (LANUV) of the state of North Rhine-Westphalia, 2-Round Biotin-aRNA Amplification Kit 3.0 (Epicentre, Madi- Germany (permit number: 87-51.04.2010.A160). Every effort son, Wisconsin) was used to generate biotin-labelled cRNA.
was made to preserve animal welfare during the prolonged Similar to other studies, cDNA was synthesized using mRNA- period of time necessary to complete the study.
specific poly(dT) primers binding the poly(A) tail of mRNA). It is possible that Oocyte collection amplification with poly(dT) primers distorts the representationof the original mRNA population by amplifying preferentially Germinal vesicle (GV) B6C3F1 oocytes were collected from the those mRNAs that have a longer poly(A) tail instead of those ovarian follicles 48 h after injection of pregnant mare serum mRNAs with short or no poly(A) tail gonadotrophin (Intergonan, Intervet, Unterschleißheim, However, the use of poly(dT) primers ensures that only those Germany). Metaphase II (MII) oocytes were collected from the transcripts are amplified that are amenable to translation at the oviductal ampullae 14 h after the injection of human chorionic time of sampling.
gonadotrophin (Ovogest, Intervet), as described (Oocytes used were ovulated from 3 week (pubertal), 8G1week (mature) and 58 Isotopic labelling G10 week (climacteric) old B6C3F1 mice.
F9 EC cells were grown for several passages in RPMI 1640 Karyotype analysis medium (PAA, Co¨lbe, Germany), supplemented with 10%dialysed FCS (Sigma), heavy amino acids 13C15 6 N2-L-lysine (K8) MII oocytes were processed for chromosome counting by 6 N4-L-arginine (R10; Silantes, Martinsried, Germany) air-drying and Hoechst dye staining, as described and glutamine and the antibiotics penicillin and streptomycin . As chromosome spreads with fewer than 20 chromosomes (Gibco, Life Technologies, Darmstadt, Germany). Labelling may be the result of chromosome loss during the spreading efficiency was examined by small-scale in-solution digests of procedure, only supernumerary chromosomes are considered the heavy labelled F9 cells and analysed by LC-MS/MS (see reliable. Therefore, the degree of aneuploidy is measured by the below). Labelled EC cells were used as a spike-in standard ratio of chromosome spreads with O20 chromosomes over when the 100 most intense proteins showed a labelling chromosome spreads with R20 chromosomes.
efficiency O97.8% (see see sectionon given at the end of this article). In orderto quantify the labelling efficiency, light F9 cells were grown in Embryo production the same medium except for the heavy amino acids, which Embryos were produced by intracytoplasmic sperm injection were replaced by the conventional amino acids 12C14 (ICSI), somatic cell nuclear transfer (SCNT) or parthenogenesis lysine (K0) and 12C14 6 N4-L-arginine (R0; Sigma), labelled as (PA) of MII B6C3F1 oocytes and cultured in a-MEM medium, as ‘light' to discriminate them from their ‘heavy' counterparts.
described Nucleus donors were C57Bl/6Jfor ICSI and OG2 (B6;CBA-Tg(Pou5f1-EGFP)2Mnn/J) for SCNT.
Protein isolation, fractionation and MS Zona-free oocytes and F9 cells were lysed in SDS lysis buffer Confocal microscopy immunofluorescence of oocytes (4% SDS, 100 mM Tris/HCl, pH 7.5, 0.1 M dithiothreitol, total The immunofluorescence and imaging protocol was described volume 70 ml), as described (We used the previously (The primary antibodies spike-in SILAC technology combined with high accuracy MS purchased from Santa Cruz Biotechnology (Heidelberg, ) to quantify age-related proteome changes Germany) were anti-STAG1 (sc-54466), anti-OCT4 (sc-8628), in mouse oocytes.
anti-ZAR1 (sc-55994) and anti-BUB1 (sc-18286). Appropriate The spike-in method employs an independently prepared Alexa Fluor-tagged secondary antibodies (Invitrogen, Life internal or spike-in standard (in our case the F9 EC cells, heavy) Technologies, Darmstadt, Germany) were matched to the against which the test proteome (the oocyte, light) is measured primaries (Alexa Fluor 488 Donkey Anti-Mouse, A-21202; after mixing them in a 1:1 ratio. The mixtures were processed Alexa Fluor 568 Donkey Anti-Goat, A-11057; Alexa Fluor 647 by the FASP–SAX procedure Donkey Anti-Rabbit, A-31573). The fluorescence signal The peptide fractions obtained from the FASP–SAX procedure intensity was quantified using Image J (U.S. National Institutes were analysed by LC-MS/MS on a LTQ Orbitrap Velos of Health, Bethesda, MD, USA).
mass spectrometer (Thermo Scientific, Waltham, MA, USA), Reproduction (2014) 148 55–72 C Schwarzer, M Siatkowski and others equipped with an Easy nano-LC system and a nano electrospray Accuracy and reproducibility of our SILAC approach source (both from Proxeon, Odense, Denmark). Briefly, We applied two controls to validate our spike-in SILAC peptides were separated by reversed-phase chromatography approach. These controls are based on the ‘ratio-of-ratios' on fused silica capillary chromatography columns (15 cm method (In the first control, we examined length, ID 75 mm; New Objectives, Inc., Woburn, MA, USA) the heavy:light ratio distribution of F9 cells and the heavy:light that were packed in-house with Reprosil pure C18 material ratio distributions of F9 cell/oocyte mixtures. The latter were (3 mm; Dr Maisch, Ammerbuch, Germany). We used gradients markedly broader than the former (judged by the number of from 2 to 27% of buffer B (80% acetonitrile and 0.5% acetic standard deviations, S.D., into which 90% of the values fall; see acid) for the SAX flow through fractions, for the pH 11 washing below), confirming the expected larger dissimilarity of the F9/ step and for the first elution at pH 8; from 5 to 35% B for SAX oocyte mixture over the F9/F9 mixture ( elution steps at pH 6 and 5; and from 8 to 40% B for the SAX In the second control, we examined the ‘ratio-of-ratios' elution steps at pH 4 and 3. Each gradient lasted 190 min and distributions of the samples of pubertal and climacteric oocytes was followed by a gradient over 10 min to 90% B and further relative to mature age oocytes and of mature age replicates elution at 90% B for 5 min before the column was equilibrated against each other. To this end, we harvested one pool of 1400 again with starting buffer A (0.5% acetic acid). The mass oocytes from mature mice (collected from 47 mice all on the spectrometer was operated in data-dependent mode (positive same day; note that the number of climacteric mice needed for ion mode, source voltage 2.1 kV), automatically switching an equivalent experiment would be around six times larger) between a survey scan (mass range m/zZ350–1650, target and split it into two mature age replicates. The protein data valueZ1!106; resolution RZ60 K; lock mass set to back- associated with the mature age replicates are available from the ground ion 445.120025) and MS/MS acquisition of the 15 most PRIDE archive. We provide the interval (number of S.D from the intense peaks by collision-induced dissociation in the ion trap zero centre) into which 90% of the ratio values fall as a measure (isolation width m/zZ2.0; normalized collision energy 35%; of how broad or narrow the distributions are. These are 9.957 dynamic exclusion enabled with repeat count 1, repeat S.D. for the pubertal vs mature oocytes, 10.189 S.D. for the duration 30.0, exclusion list size 500 and exclusion duration climacteric vs mature oocytes, but 4.286 S.D. for the mature age set to 90 s; double charge and higher charges were allowed).
replicates (These resultsindicate that diversity between the different age groups is larger Protein identification and quantification than within the same age group – a prerequisite to analyse theoocytes of different maternal ages comparatively. Unless Raw data were processed by MaxQuant Software (v, otherwise stated, the mature age group was used as a reference Martinsried, Bavaria, Germany) involving the built-in Andro- to analyse the variation and concordance of protein and meda search engine. The search was performed against the transcript levels during maternal ageing.
International Protein Index database (mouse IPI version 3.73;concatenated with reversed sequenceversions of all entries and supplemented with common Bioinformatics and statistical data analysis contaminants. Parameters defined for the search were trypsin The microarray raw data were imported into the R environment as the digesting enzyme, allowing two missed cleavages; a (We quantile-normalized the data and minimum length of six amino acids; carbamidomethylation at filtered out probe sets of a low signal level using the cysteine residues as fixed modification, oxidation at methion- Agi4x44Preprocess package (Bioconductor; ine and protein N-terminal acetylation as variable modifi- ) with default parameters. Expression values of probe sets cations. The maximum allowed mass deviation was 20 ppm for mapping to the same gene were averaged, yielding 22 334 the MS and 0.5 Da for the MS/MS scans.
unique entities. The correlation of mRNA and protein Protein groups were identified with a false discovery rate set abundance values was calculated by Kendall rank (non- to 1% for all peptide and protein identifications separately, parametric) t correlation coefficients.
when there were at least two matching peptides, at least one of In the proteome data provided by the MaxQuant Mass which was unique to the protein group. Primary quantification Spectrometer Data Analysis Software, proteins are placed in was performed using the heavy F9 lysate mix as an internal one group if the identified peptide set of one protein is equal to standard, and ratios between corresponding heavy and light or contained in a second protein's peptide set. Then, the peptide versions were normalized to this mix and expressed as proteins with the highest peptide count in a group are retained.
H/L (i.e. heavy/light SILAC internal standard/sample). All these The IPI (updated to release 3.87) identifiers of these proteins primary protein ratios are the means of at least two (heavy and were mapped to ENTREZ identifiers with an existing MGI light) peptide ratios. Direct comparison of oocyte samples was symbol. We obtained a single ENTREZ identifier for 96.22% of achieved by a ‘ratio-of-ratios' calculation, which is possible the protein groups to which we assigned the logarithm (to base 2) because the internal standard was the same for all samples of the heavy:light ratio calculated by MaxQuant. Ties that may occur due to isoforms (splice variants) or close paralogs that are The MS proteomics data have been deposited into the not distinguishable based on the peptides identified are ProteomeXchange Consortium ( handled by downweighting the logarithmic heavy:light ratio, via the PRIDE partner repository ( defining the base b of the logarithm as bZ2C(TK1)/T, where with the dataset identifier PXD000512.
T is the number of ties. Thus, proteins with ties have lower Reproduction (2014) 148 55–72 SILAC proteome analysis of mouse oocyte ageing influence in subsequent analyses. In the case of proteins we refer to this algorithm as the ‘elimination mapping to the same ENTREZ identifier, we average the algorithm'. Finally, GO term heat maps were generated with a logarithmic heavy:light ratios. We subjected the proteins P value cut-off of 0.01, and a gray gradient corresponding to the identified to further analysis using a ‘ratio-of-ratios' ( P value from light gray (slight overrepresentation, high P value) ) describing the ‘age-transition-based expression to dark gray (marked overrepresentation, low P value), using change' from pubertal to mature and from mature to white as default (no overrepresentation).
climacteric. Specifically, we subtracted the log2 F9:oocyteratio of the mature age group from the log2 F9:oocyte ratioof the pubertal group, and we subtracted the log2 F9:oocyte ratio of the climacteric group from the log2 F9:oocyte ratioof the mature age group.
Maternal ageing of oocytes features both deteriorationand improvement of functional properties Correlation analysis of transcriptome and proteome In order to give our proteomic analysis a phenotypicfoundation, we firstly obtained a functional portrait We asked whether maternal ageing within the proteome– of the MII oocytes ovulated by B6C3F1 mice aged transcriptome intersection causes regulated changes in the 3 weeks (pubertal), 8 proteome of oocytes compared to their transcriptome. We G1 weeks (mature) and 58G10 considered that the analytical methods of LC-MS/MS and weeks (climacteric). As expected, the number of oocytes microarray may have different dynamic ranges; we, therefore, ovulated decreased (and the frequency of calculated their correlation using a statistical method that is hyperhaploid karyotypes increased during maternal independent of the absolute amount of change. The Kendall ageing: 7% (1/13), 0% (0/39) and 30% (9/30) respec- tau (t) coefficient of correlation measures the similarity of the tively. In spite of these negative changes, the proportion proteome (X) and transcriptome (Y) protein/gene orderings of oocytes competent for blastocyst formation increased when ranked by their respective changes of abundance, with maternal age, as measured by blastocyst rates at regardless of any threshold. In other words, it measures the 96 h after ICSI ( similarity of the orderings of the data (X and Y) when ranked Higher cell numbers accompanied the increase in by each of the quantities. If the agreement between the two blastocyst rates. The blastocysts' total cell numbers rankings is perfect (i.e. the two rankings are the same), the in the ICSI group increased significantly with maternal coefficient has a value of 1. If the disagreement between the age (pubertal: 36.7G13.6 cells, nZ32; mature: 45.8G two rankings is perfect (i.e. one ranking is the reverse 15.7 cells, nZ33; climacteric 55.8G14.2 cells, nZ16; of the other), the coefficient has a value of K1. If X and Y ANOVA, PZ0.0001; multiple comparisons using t-test, are independent, then the coefficient is zero.
In addition to fertilization (ICSI), developmental competence was also tested by parthenogenesis (PA) and GO BP overrepresentation analysis by SCNT from cumulus cells (B). PA lacks In order to address the concern of a sampling bias of the contribution from sperm; SCNT eliminates a possible proteome with respect to the transcriptome, we sampled 1000 effect of meiotic aneuploidy on development. In both random sets from the transcriptome, of the same size as the PA and SCNT, maternal age correlated positively with proteome, and calculated the GO semantic similarity score an increase in blastocyst rate at 96 h (Fisher's exact test, between the proteome and the transcriptome and between the P!0.01). Cumulus cells from a transgenic donor harbour- random sets and the transcriptome using the GOSemSim ing green fluorescent protein (GFP) under the control of the regulatory element of Pou5f1 i.e. Oct4 (Oct4–GFP) Considering both proteome age transitions, two gene lists were used for SCNT, which allows the visualization of ranked by the expression values were subjected to GO BP pluripotent gene expression. Higher OCT4–GFP intensity, overrepresentation analysis. Following which is a predictor of developmental competence we applied a cut-off-free protocol to each gene list, (was observed in derivative cloned following rank square transformation, treating positive and morulae obtained from the oocytes of older mothers negative expression changes in the same way. We then calculated the Mann–Whitney U test statistics for all gene Z0.0364; multiple comparisons using t-test, sets that correspond to GO BP terms, investigating whether PZ0.011 for climacteric vs pubertal age; .
genes in these gene sets rank higher in the transformed ranking Although it is well established that deterioration of than the genes not in the gene set. Then, we drew 1000 random genome stability (e.g. meiotic aneuploidy) is a hallmark sets of the same size from all genes in the detected proteome, of oocyte ageing, our data attest to the presence of other for each GO BP term. Again using the transformed ranking, we features that clearly do not deteriorate with maternal calculated the Mann–Whitney U test statistics for the random ageing, as revealed by improved blastocyst formation, sets, thus obtaining P values describing the significance of the blastocyst cell numbers and Oct4 transgene expression.
GO term enrichment. We also applied an algorithm in Our aim here is to shed light on how oocytes age in processing the ontology terms that returns more specific GO molecular terms. The variable rates and qualities of BP terms by considering the dependencies between them blastocyst formation document that the nature of the Reproduction (2014) 148 55–72

C Schwarzer, M Siatkowski and others OCT4–GFP intensity Number of oocytes o Developmental rates to blastocyst obtained from pubertal, mature age and climacteric oocytes after SCNT, ICSI and PA.
28 (9.2±3.4)†* 63 (19.1±32.3)†* 156 (32.8±31.1)†* 160 (38.6±4.2)†* 132 (44.6±14.8)†* 166 (69.5±7.4)†* Climacteric (58±10) 126 (48.8±5.4)†* 95 (72.0±17.5)†* †Fisher's exact test, pairwise comparison related to the one-cell stage and within groups (e.g. SCNT), *P<0.01.
n, number of replicates.
Figure 1 Increased developmental potential of oocytes ovulated from B6C3F1 mice of advanced maternal age. (A) Numbers of oocytes ovulated inassociation with blastocyst rates after ICSI in pubertal, mature and climacteric mice. Significance was tested by ANOVA, pairwise comparison wasmade with a multiple comparison t-test (*P%0.05). (B) Blastocyst rates observed after fertilization (ICSI), parthenogenesis (PA) and cloning (SCNT)given as both total (sum over replicates) and meanGS.D. Significant differences in blastocyst rates between age groups were compared using pairwiseFisher's exact test. (C) Bright field and fluorescent images of morulae after SCNT from OG2F1 cumulus cells (transgenic for Oct4-GFP); size bar,50 mm. Significance was tested by ANOVA (*P%0.05).
starting material – the oocytes – was not constant during respectively 140 female mice were required in the maternal ageing, beyond the notorious phenotype of climacteric group to obtain 700 oocytes. Given the meiotic aneuploidy. The features of maternal ageing may scarce amount of material, especially in the climacteric be rooted in the nucleus or in the cytoplasm of the group, we adopted a spike-in (SILAC) quantitative oocyte. While these cellular components are physically method for proteome analysis, coupled to high- mixed in MII oocytes (absence of nuclear envelope), resolution liquid chromatography combined with MS they can be disentangled in silico by performing GO (LC-MS/MS). We used F9 cells labelled isotopically with analysis of gene expression data in the domain ‘cellular heavy lysine (Lys8) and heavy arginine (Arg10) as component'. We considered that since proteins relate material for the spike-in. Oocyte lysates were mixed more closely to cell phenotypes than mRNAs, protein 1:1 to lysates of F9 cells that had been labelled efficiently analysis may reveal aspects of the maternal age effect (97.8%; Thus, the proteins that could not be grasped so far by transcript studies.
detected in both oocytes and F9 cells are quantifiablerelative to each other in this study. Using a linear ion-traporbitrap hybrid mass spectrometer (LTQ Orbitrap) and A SILAC screen of the proteome of MII mouse oocytes the MaxQuant Proteomic Software, we were able to of different maternal age identify a total of 3268 different protein groups, of which Seven hundred zona-denuded MII oocytes were used for 2654, 2639 and 2617 groups were found in pubertal, each of the three age groups (3 weeks, pubertal; 8G1 mature and climacteric MII oocytes respectively weeks, mature age and 58G10 weeks, climacteric; see (pubertal h mature, 2450 groups and mature h for a graphical visualization of the SILAC climacteric, 2451 groups). Here, each protein group experimental set-up). While 15 and 24 female mice identified is defined by an F9:oocyte ratio (heavy:light were sufficient in the pubertal and mature age groups isotopic ratio). Accuracy and reproducibility of our SILAC Reproduction (2014) 148 55–72

SILAC proteome analysis of mouse oocyte ageing approach were validated (see that may occur because of isoforms/splice variants (see and ‘Materials and methods' section).
‘Materials and methods' section for details), mapping The protein groups identified were mapped to ENTREZ resulted in 2773 gene identities for the pubertal age identifiers with an existing MGI symbol. Owing to ties group, 2757 for the mature and 2732 for the climacteric with Lys8 and Arg10 Non-labelled oocyte samples of different age groups (light) ratio F9:oocyte (heavy/light)= quantified protein level 2 F9:oocyte (pubertal) 2 F9:oocyte (mature age) Log2 F9:oocyte (climacteric) F9 oocyte (pubertal) F9 oocyte (climacteric) Age transition-based expression change: F9 oocyte (mature age) F9 oocyte (mature age) => Proteins changing during first transition => Proteins changing during second transition Simultaneously detected (8±1 weeks) (58±10 weeks) in all age groups No. of transcripts (microarray) No. of detected protein groups No. of corresponding gene identities (proteins) – without probe on microarray: 140– with corresponding transcripts below detection limit: No. of gene identities in No. of gene identities (proteins) detected in age transitions (first age transition) (independent of transcript detection) (second age transition) No. of gene identities (proteins) changingmore than four fold in age transitions (independent of transcript detection) (first age transition) (second age transition) (in both age transitions) Figure 2 Schematic overview of the SILAC workflow and summary. (A) Oocyte lysates were each mixed 1:1 with labelled F9 cell lysate (SILACreference standard, ‘spike-in'). Mixtures were fractionated and analysed by LC-MS/MS. Primary quantification was performed using the heavyF9 cells as SILAC internal standard (*): ratios between corresponding heavy (F9 cells) and light (oocyte) peptide versions were used to determineprotein expression levels (expressed as H/L, heavy/light, i.e. SILAC internal standard/sample). Secondary quantification (age transition-basedexpression change) was performed by dividing the individual H:L peptide ratios of the age groups to be compared (a ‘ratio-of-ratios' calculation).
(B) Summary of the amounts of transcripts/proteins detected, identified and analysed under the specified criteria.
Reproduction (2014) 148 55–72 C Schwarzer, M Siatkowski and others age group. A summary of all proteins identified and all and from mature to climacteric (second age transition), peptides detected with their mass accuracies is provided using mature as the point of reference. We adopted a in see section on fourfold threshold to set significance, because it is in a given at the end of this article.
large excess of the variation (max. 0.41-fold) shown by Representative protein spectra are shown in housekeeping gene products HPRT1, H2AFZ and PPIA A total of 2324 of these mapped gene (A) and because twofold identities were common to the three age groups (B fluctuations in proteomics can also be seen in technical and The raw protein data replicates (Of the 2066 proteins associated with the three age groups are available at shared by all age groups and with detected mRNA, 48 the ProteomeXchange Consortium (see link in ‘Materials and 42 proteins varied greater than or equal to fourfold in and methods' section).
the first and in the second age transition respectively. The We performed conventional microarray (Agilent, mathematical union (g) of these two groups of proteins Santa Clara, California) analysis of these oocytes in yields 69 proteins (3% of 2066). In contrast to proteins, parallel with the proteome analysis. The transcriptomes the abundance of only one mRNA, namely Cenp-e, were generated from 20 MII oocytes collected in changed fourfold during the first age transition and none biological triplicates for each age group (pubertal, in the second. CENP-E is a kinesin-like protein mature and climacteric). A total of 22 334 gene identities associated with kinetochores. Its mRNA was identified were common to the transcriptomes of the three age as age-regulated in mouse oocytes groups (The raw microarray data associated Overall, the quantitative change in the oocyte with this manuscript are available at the Gene Expression proteome (69/2066 proteins; 3%) exceeded the quan- Omnibus (see link in ‘Materials and methods' section).
titative change in the protein–matched transcriptome A 140 of the 2324 gene identities of the proteome had (1/2066 genes 0.05%) (c2 test, PZ2.46!10K16), and the no corresponding probe on the microarray ( concordance of the two is essentially nil. The overall lack of correlation between proteome and transcriptome, given at the end of this article); cognate mRNA was calculated on all 2066 genes/proteins, is summarized below detection level for another 118 gene identities, by a Kendall t rank correlation coefficient that is not despite the presence of the probe on the microarray significantly different from zero: 0.0262 in the first ). Hence, 2066 gene identities age transition (PZ0.0743) and 0.0053 in the second were simultaneously detected in the transcriptome and age transition (PZ0.7184, see also ‘Materials and proteome in all age groups (transcriptome–proteome methods' section).
intersection, A summary of thenumbers of proteins detected is provided in Bioinformatics analysis based on GO terms revealed The most differently expressed proteins of the ageing that the two datasets for the proteome and the oocyte proteome are predominantly localized in thenucleus transcriptome feature similar BPs. The GO semanticsimilarity (GOSemSim) score is 0.899, which is within Because quantitative change (greater than or equal to the interval defined by the lower and upper quartiles fourfold) in the oocyte transcriptome is not correlated (0.894 and 0.901 respectively) of the distribution of with that in the proteome, we decided to consider all the similarity scores between the random sets and the proteins detected in subsequent analyses, irrespective of transcriptome (see ‘Materials and methods' section for whether they have cognate mRNAs or not. This decision details on how the GOSemSim was calculated).
resulted in a slight increase in the number of changing Although the number of proteins identified is smaller (greater than or equal to fourfold) proteins, from 48 to 55 than the number of transcripts, the proteins we detected in the first age transition and from 42 to 49 in the second can thus be considered representative of the complete age transition. These proteins are listed in along with the direction (b, increase and a, decrease) ofquantitative change and the GO domain ‘cellularcomponent' with which they are annotated. The two Quantitative change in the oocyte proteome is not age transitions (first, puberty to mature age and second, predictable from cognate mRNAs mature age to climacterium) will be contrasted with each The protein analysis conducted so far has been other in order to identify those protein changes that are qualitative (presence/absence). In order to appreciate specific to the old age.
the quantitative impact of maternal ageing on the oocyte The relationship of the 55 and 49 age-regulated proteome, we analysed protein abundances, as defined proteins (to classic features of ageing is by the ‘ratio-of-ratios' (; see described hereafter. Roles in spindle assembly, mainten- and ‘Materials and methods' section). This reflects age- ance of centrosome integrity and chromosome segre- transition-based expression changes, which describe the gation are featured by HAUS7, ACTR1B, BUB1 and TNT.
transition from pubertal to mature (first age transition) Another classic feature of ageing, namely oxidative Reproduction (2014) 148 55–72 SILAC proteome analysis of mouse oocyte ageing First age transition Second age transition Figure 3 Expression profiles of selected proteins in B6C3F1 oocytes during maternal ageing. Change in protein abundances of (A) housekeepers,(B) factors related to oxidative stress, (C) culprits of ageing revealed by transcriptome studies, (D) structural maintenance of chromosomes (SMC) andspindle assembly checkpoints (SACs) and (E) maternal-effect factors. Values are expressed as fold-change during first and second age transition.
Reproduction (2014) 148 55–72 C Schwarzer, M Siatkowski and others Table 1 Proteins undergoing change in expression during the first (puberty to mature age) and second age transition (mature age to climacteric) inexcess of fourfold (in alphabetical order).
First age transition Second age transition Cellular component Cellular component C, cytoplasmic; N, nuclear; NA, not annotated.
aProteins changing in excess of fourfold in both age transitions.
stress, is represented in SART1 and ERO1LB. These Other age-regulated proteins have a relationship to two proteins are very loosely related to oxidative oocyte quality and include ZAR1 (maternal-effect stress according to the database of the European factor), PAPOLA (mRNA maturation) and TCL1B1 Bioinformatics Institute (and (developmental potential). PAPOLA is the poly(A) they are not known to play a role in mouse oocytes.
polymerase a, which adds adenosine residues to create Reproduction (2014) 148 55–72 SILAC proteome analysis of mouse oocyte ageing the 30-poly(A) tail of mRNAs (). TCL1B1 climacteric oocytes. We have tested and confirmed the is associated with developmental potential in early STAG1 profile in situ – directly in MII oocytes of the three embryos Other proteins that age groups – using confocal immunofluorescence changed in abundance include members related to microscopy and compared the measured intensities.
cytoskeleton assembly/organization (e.g. TMS4BX, The antibody results matched the LC-MS/MS results ACTR1B, TUBA3A, MYL1, PTK7, EML4, PLEC, PTK2 (and A0; pubertal vs mature age, P!0.0001; and JUP) and to protein modification (e.g. CRNKL1, climacteric vs mature age, P!0.0001; Dunnett's test).
While the abundance of MAD2L1 and other proteins Inspection of the GO domain ‘cellular component' of spindle assembly and chromosome segregation varied reveals that the proteins in the second age transition that less than twofold, another component of the spindle increased with ageing are mainly cytoplasmic, while checkpoint, BUB1 (), varied those that decreased are mainly nuclear ; c2 test, more than sevenfold (As with STAG1, BUB1 PZ0.017). This allocation is not significantly different abundance was tested by immunofluorescence analysis from 1:1 in the first age transition (c2 test, PZ0.375).
(B and B0), but it could not be confirmed. This Among the 49 proteins changing in the second age discrepancy between SILAC and immunofluorescence transition, 24 also change in the first transition, while 25 data is discussed below.
changed only in the second age transition. Reflecting A maternal-effect factor, ZAR1, is among the most findings already described above, these 25 proteins have differently expressed proteins of the ageing oocyte a prevalence of cytoplasmic terms among the increasing proteome (We also examined the other proteins and of nuclear terms among the decreasing maternal-effect factors. Of the 27 maternal-effect proteins (c2 test, PZ0.054). PAPOLA and ZAR1 are proteins known (17 were detected in among the latter ().
oocytes of all three age groups MATER (NLRP5),the maternal antigen that embryos require, is para-digmatic of the maternal-effect factors Proteins of known candidate genes in the ageing oocyte and was detected together with the three other members of the subcortical maternal complex (SCMC), namely We were surprised in our threshold-based analysis OOEP (FLOPED), KHDC3 (FILIA) and TLE6 ( (that, except for BUB1, the culprits of oocyte Abundance of these factors was stable during ageing were missing, such as BCL2 and BAX (mito- maternal ageing, as was the abundance of other chondrial function and apoptosis; prominent maternal-effect factors, e.g. STELLA (DPPA3) APACD, SOD1, TXN1 (oxidative stress; and OCT4, but not ZAR1. ZAR1 abundance declined in oocytes of climacteric females, as confirmed in situ using confocal immunofluorescence and comparison of the NUMA1, SMC1B (spindle assembly and chromosome measured intensities (and C0; Dunnett's test for integrity/stability; , ZAR1 pubertal vs mature age, PZ0.6108 and climac- and DMAP1, DNMT1, DNMT3A, teric vs mature age, PZ0.0043). OCT4 abundance HDAC1/2 (epigenetic modification; , increased slightly in oocytes of climacteric females, but ). Therefore, we searched for them immunofluorescence analysis revealed a small yet directly, disregarding the fourfold threshold. APACD, significant decline in OCT4 levels (Dunnett's test, BRCA1, DMAP1 and SMC1B were not detected among P%0.0218; D and D0). These discrepancies the candidates, while the ones that were detected varied between SILAC and immunofluorescence data are less than twofold, except BAX, HDAC1 (and discussed below.
BUB1 In contrast to these known culprits ofoocyte ageing, three as yet uncharacterized proteins Molecular signature of oocytes during maternal ageing loosely related to oxidative stress were found to changein excess of twofold, namely ERO1LB, SART1 and The most differently expressed proteins of the ageing glutathione S-transferase omega 1 (GSTO1; B).
oocyte proteome were, only in part, those we had We continued our analysis focusing on the proteins expected based on mRNA studies published. In previous responsible for genome stability (e.g. ploidy) and transcriptome studies, only a small number of GO BP embryonic genome activation (e.g. maternal-effect were affected in old oocytes, based on thresholds that factors). While SMC1B was not detected in our study, were sometimes as low as 1.4- to 1.5-fold ( two other members (SMC1A and SMC3) of the structural maintenance of chromosomes (SMC) complex, which set out to perform GO analysis to discover the signature holds the sister chromatids together, were detected in of the oocyte proteome during maternal ageing.
oocytes of all three age groups, but their abundance We performed GO overrepresentation analysis using varied less than twofold D). By contrast, the SMC ranks instead of age-transition-based expression changes cofactor STAG1 decreases in excess of twofold in so that the conclusions of the GO analysis are valid Reproduction (2014) 148 55–72

C Schwarzer, M Siatkowski and others Intensity (arbitr Intensity (arbitr Intensity (arbitr Intensity (arbitr Figure 4 SILAC abundance of selected oocyte proteins validated in situ by immunofluorescence. (A, B, C and D) Representative pictures ofimmunofluorescence staining of STAG1, BUB1, ZAR1 and OCT4 on GV and MII stage oocytes, with DNA counterstaining (YOPRO-1, Life Technologies,Darmstadt, Germany). (A0, B0, C0 and D0) Results of the quantification of the immunofluorescence signal in MII oocytes, performed with Image-J Software.
Significance was tested by ANOVA, pairwise comparison was made with the Dunnett's test, taking the mature age as reference (*P%0.05).
irrespective of any threshold that we may set, such as fourfold (see ‘Bioinformatics and statistical data analysis' Although the oocytes ovulated by older mice complied section for details). We further applied the ‘elimination with the expected reduction in number and increase of algorithm' () to account for redundant meiotic aneuploidy, other commonly expected changes GO terms. We considered all the proteins of each age could not be verified. Oocytes of older mice, for transition (2450 and 2451 respectively), independent of example, were superior to younger counterparts at any threshold and of the detection of the protein itself in accumulating cells during cleavage, forming blastocysts the other age transition (and and expressing Oct4–GFP, irrespective of the develop- see section on given mental stimulus (ICSI, PA and SCNT). These obser- at the end of this article). Our overrepresentation vations confirm and extend our previous study ( analysis in GO BP terms shows that the two age Thus, while the well-known fact of maternal transitions feature different BPs (the larger the age-dependent deterioration of oocyte ploidy was enrichment, the more the colour shift to dark gray).
confirmed, similar deterioration is not applicable to all During the first age transition, processes related to the properties of oocytes, some of which are preserved or ‘regulation of blood pressure', ‘stem cell maintenance' even improved. The blastocyst phenotypes are evidence that the nature of the starting material – the oocytes – is overrepresented. During the second age transition, not constant during maternal ageing and that there is however, the analysis shows overrepresentation of BPs more to oocyte ageing in vivo than the notorious associated with ‘RNA processing', ‘mRNA splicing, via increase of aneuploidy. Using the innovative proteomic spliceosome', ‘positive regulation of nucleocytoplasmic tool to appreciate this complexity, we are going to transport' and ‘heart morphogenesis' (), among discuss that i) proteome analysis of mouse oocytes may others. In line with the most differently expressed not be surrogated with transcriptome analysis and ii) the proteins, these are not the BPs one would expect based classic features of ageing may not be transposed from on the GO analyses of transcriptome data, which feature somatic tissues to oocytes in a one-to-one fashion.
terms related to, for example, oxidative damage and So far, the molecular bases of oocyte ageing, e.g.
stress response ( meiotic aneuploidy, have been searched for in the ) and inflammation ( mRNA world. Expectations from mRNAs are now This discrepancy is also discussed below.
confronted with unprecedented information gained Reproduction (2014) 148 55–72

SILAC proteome analysis of mouse oocyte ageing Regulation of blood pressure Stem cell maintenance Cellular response to light stimulus Transmembrane receptor protein tyrosine kinase signaling pathway Peptidyl tyrosine phosphorylation mRNA splicing, via spliceosome Positive regulation of nucleocytoplasmic transport Heart morphogensis first age transition second age transition High P value Low P value First age transition: pubertal to mature ageSecond age transition: mature age to climacteric Figure 5 Gene ontology (GO) overrepresentation analysis reveals proteomic signature of oocyte ageing. Overrepresentation heatmap of GO BPterms of 2450 and 2451 proteins detected in the first and second age transition respectively. For the heatmap, the two protein lists were ranked byexpression values. Rank square transformation was applied for threshold-free and direction of change (up and down)-independent ranking.
Random testing using Mann–Whitney U statistics was applied to obtain P values (cut-off 0.01). Colour code: gray gradient corresponding to theP value, from light gray (light overrepresentation, high P value) to dark gray (marked overrepresentation, low P value), using white as default(no overrepresentation).
from proteomic analysis, providing us with unexpected proteome correlation also has been described to be poor results. Our study revealed the highly dynamic character when analysing stably growing cell lines that are in a of the oocyte proteome during maternal ageing in vivo, steady-state and we deem it, therefore, very unlikely that whereas proteome and transcriptome were qualitatively our correlation analyses have been hampered by similar in GO semantic composition. Overall, there was technical matters (). Further, it is not a higher prevalence of proteins changing in excess of surprising that protein fluctuations are inherently larger fourfold compared to transcripts (69/2066 vs 1/2066) than mRNA fluctuations. and minimal concordance between changes of the 2066 observed that distributions of protein copy number per pairs of protein and transcript values (Kendall t close to cell have two to three times (in log10 scale) higher zero). It should be noted that poly(dT)-primers have been median and higher variation compared with mRNA.
used for the required pre-amplification step when Whether these larger fluctuations shed light on the pursuing oocyte microarray analyses. Therefore, only maternal age effect in oocytes is crucial.
mRNAs with a poly(A) tail that could effectively be Climacteric oocytes being hard to come by (140 mice translated at the time of sampling, i.e. the MII stage needed for 700 oocytes), we relinquished the replicates oocyte, have been analysed. This may explain why and the conventional approach of setting twofold certain proteins have been detected for which no thresholds combined with P values in order to make corresponding mRNA could be identified. It is reason- the call of significance. Instead, we combined a spike-in able to assume that these RNAs had existed during method (SILAC internal reference) with a higher oocyte maturation but lost their poly(A) tail at the time of threshold of fourfold change. This threshold is higher ovulation to mark them for degradation. This fact may than the difference observed by also have influence on the mRNA expression levels in their comparison of ageing in different somatic tissues detected in MII oocytes in general and may partially and substantially higher (about ten times) than the explain the missing correlation between transcriptome variation observed in housekeeping proteins (0.4-fold and proteome in oocytes. However, the transcriptome to variation), which are overall stable. Searching the oocyte Reproduction (2014) 148 55–72 C Schwarzer, M Siatkowski and others proteome for a group of changes (e.g. GO terms) that blastocyst stage as the developmental endpoint. This would stand out from the background, our SILAC study confirms and extends our previous study ( revealed several proteins that account for the malfunc- Although we should not aim to explain the tion of the chromosomal apparatus, but few that reflect developmental performances on the basis of the oxidative stress or damage. As this study is the first of proteome detected (which is still incomplete, despite its kind, one has to be careful not to over-interpret data.
the best analytical technology used), we note an Yet, we want to discuss some individual genes that interesting age-dependent change in the abundance of specifically caught our attention. Our SILAC results two factors that may facilitate, or impede, development.
revealed change in the abundance of the SMC cofactor The abundance of GSTO1 increased in both the first and STAG1, as well as in that of the SAC protein BUB1 and second age transition. Although there is no study of the histone deacetylase protein HDAC1. This latter GSTO1 in mammalian oocytes, impaired synthesis of protein modulates the ability of chromosomes to interact glutathione is indicted as the main cause for compro- with spindle microtubules and is depleted in climacteric mised developmental potential of pubertal mouse oocytes, thereby jeopardizing proper chromosome oocytes in mice (hence, the increased segregation (Unlike STAG1 and abundance of GSTO1 in old oocytes may facilitate BUB1, our SILAC results reveal no marked change in development. Abundance of the cell death inducer BAX the abundance of the core cohesin factors SMC1A was high in pubertal oocytes and decreased in the and SMC3. It has been proposed that the efficiency first age transition. Induction of Bax gene expression of oxidative phosphorylation in the ageing oocyte is by in vitro culture conditions correlates with reduced degraded by free radical attack ( developmental rates of mouse embryos ( however, neither metabolism- nor oxidative stress- ); hence, the higher abundance of BAX in associated genes featured substantial change of pubertal oocytes may explain, at least in part, their lower abundance in our quantitative proteome data. A change developmental competence compared with mature-age of proteins very loosely associated with response to and climacteric oocytes. Certainly, the possibilities are oxidative stress, e.g. SART1 and ERO1LB, was detected not exhausted here, and there may be additional factors in the first and in the second age transition; however, whose accumulation (developmental agonists) or a role of these two proteins has not been described in depletion (developmental antagonists) in oocytes could oocytes as of yet.
explain the increase in developmental potential How reliable are fold-differences of protein abun- observed during maternal ageing.
dance in the SILAC measurement? We found a similar Although the proportion of blastocyst formation- abundance trend between our SILAC data and immuno- competent oocytes increased with maternal age, fluorescence analysis for STAG1, but not for BUB1. We aneuploidy would impair subsequent development.
also examined ‘maternal effect factors', i.e. ZAR1 and Follow-up of the embryos into post-implantation OCT4 (), and we confirmed ZAR1, but not development is beyond the scope of this study and OCT4. On the one hand, while antibodies are limited in would not add to what is known already their ability to detect partially degraded proteins due to The higher blastocyst potential of older oocytes, the lack of proper 3D structure, proteomics is only however, does suggest that other features of ageing dependent on small stretches of peptide sequence and, oocytes (including non-nuclear processes) may be less therefore, a degraded protein might still be detected by affected by maternal ageing than chromosome stability.
LC-MS/MS and MaxQuant. In this respect, while The majority of proteins that underwent quantitative proteomics offers more technical accuracy, immuno- change in the second age transition are classified as fluorescence offers more biological relevance. On the cytoplasmic in the GO domain ‘cellular compartment'.
other hand, there are many antibodies that detect distinct Among the nuclear proteins not directly associated with protein isoforms (e.g. OCT4A only), whereas LC-MS/MS SMC, our data revealed a reduced abundance of the can, in principle, detect all isoforms (contingent on the poly(A) polymerase a (PAPOLA) and HDAC1, and an quality of the database). It is important to note that increased abundance of TCL1B1 in old oocytes. As the isoforms were not considered during the SILAC analysis recruitment of many mRNAs for translation depends on in this study. The prominent maternal factor OCT4, the length of their poly(A) tail (the whose A isoform is dispensable for embryo development reduction in the abundance of PAPOLA may affect the varied less than twofold during the age translation efficiency of mRNAs and could further transitions. Although OCT4 abundance correlates exacerbate the already poor correlation between tran- positively with developmental potential, it is probably scriptome and proteome. We speculate that, at least for not a precondition, rather it is an effect ( certain cellular functions, oocytes can react to maternal ageing by operating a shift in how gene transcripts are On the functional level, we observed an increase in used, leading to the protein profiles not matching the developmental rates with ageing in an in vitro culture mRNA profiles. PAPOLA is also one of the 118 proteins environment outside of the aged female, with the with no detected cognate mRNA, together with HDAC1/2 Reproduction (2014) 148 55–72 SILAC proteome analysis of mouse oocyte ageing (member of the ‘reprogrammome'; , different pattern of gene expression changes than ageing see section on given at the end of of somatic organs.
this article; suggesting that the There is one final issue raised by our study. We have amount of enzyme for these two factors may not be characterized the maternal age-effect on oocytes, but what replenished without de novo transcription. We speculate is the underlying cause? Perhaps, the answer to this that this might be a fatal hurdle for SCNT question has to be searched in the somatic niche of the embryos because de novo transcription of the Papola ovary. The role of the niche has been investigated in the gene is dependent on prior nuclear reprogramming, testis. Spermatogonial stem cells (SSCs) do not age or age which is known to be inefficient. TCL1B1, which showed very slowly, as shown by the successful consecutive an increase in abundance in old oocytes, is a member of transplantation of SSC populations from the testes of old the T cell leukaemia/lymphoma 1 (TCL) family and is mice to the testes of young recipients. These transplanted specifically expressed in oocytes and stem cells.
SSCs are preserved long past the normal life span of the old Importantly, Tcl1 is a known downstream target of donor, for more than 3 years These Pou5f1 which is one of the four factors observations suggest that paternal infertility results from that reprogramme differentiated somatic cells to become deterioration of the somatic niche, not from deterioration pluripotent cells TCL1B1 of the germ cells. Ideally, one would like to perform a is implicated in the developmental potential of mouse similar investigation also on the ovary, to test whether the embryos ). The abundance of TCL1B1 quality of oocytes is preserved beyond the life span of the is increased almost tenfold in old oocytes, suggesting an female by transplanting her oocytes to a younger niche. In explanation for their increased blastocyst rate.
principle, this experiment is feasible When we examined the molecular signature of the ), but very challenging. Until it is performed, ageing oocyte proteome, we could not confirm the BPs we think that the available literature already hints to an indicted by previous transcriptome studies as also being effect of the niche on oocyte ageing. altered in the ageing proteome, even though the showed that adult mice maintained under 40% proteome is representative of the transcriptome as judged caloric restriction (CR) did not exhibit maternal age-related by GO semantic similarity. Transcriptome studies pointed increases in oocyte aneuploidy, chromosomal misalign- at genes involved in mitochondrial function, oxidative ment on the metaphase plate, meiotic spindle abnormal- damage and stress responses ), ities or mitochondrial dysfunction, all of which occurred in genes associated with protein folding/response to oocytes of age-matched controls that were fed ad libitum.
unfolded proteins, protein metabolism/metabolism, intra- As the bulk of oocytes have already formed at the time of cellular transport and cell cycle as well CR treatment, and as CR suppresses ovulation ( as stress response and response to oxidative damage we think that the effect of CR is unlikely to (. Although these three studies also depend on the oocytes themselves, rather on their niche.
found other families of genes responsive to maternal Could the above considerations be extrapolated to ageing, the number of genes was small and their fold- humans? Caution should always be exercised in extra- change was low (sometimes as low as 1.4- to 1.5-fold).
polating, due to species-specific differences in life span, The new technology of SILAC MS can now inform the oogenesis, duration of the ovarian cycle and onset of study of oocyte ageing. The most significant terms of BP in menopause (). The average life the oocyte proteome include ‘stem cell maintenance' in expectancy of B6C3F1 females is 128 weeks ().
the first age transition and ‘RNA processing' in the second The climacteric mice used in our study were aged up to 68 age transition. It is tempting to speculate that these terms weeks, which would correspond to women 44 years old, are consistent with a ‘memory' of the oogonium (stem assuming a linear relationship and a human life expect- cell)-to-oocyte transition (a recent event in ovaries of ancy of 84 years for females. Thus, the cells of our study 3-week-old mice) and with a change in the mode of could have been even older than 68 weeks, which is not a transcript usage at old age respectively. As the MII oocyte problem with liver, kidney and brain ( is a transcriptionally silent, terminally differentiated non- but it is a problem with the ovary as it is depleted of dividing cell, it cannot use the window of opportunity of oocytes long before the animal dies. Human oogonia S-phase to impose new epigenetic marks that alter divide more times than mouse oogonia until they enter transcription, such as cycling somatic cells; thus, it meiosis and these divisions changes its phenotype through post-transcriptional may introduce DNA mutations. Perhaps more importantly, actions. This may partly explain why the GO terms if we follow up on the effect of CR in mice, then we should commonly found enriched in ageing somatic tissues (e.g.
consider that the basal metabolic rate per gram of body inflammation) did not characterize the ageing oocyte weight is seven times greater in mice than in humans proteome. This observation is in line with the conclusion (. Therefore, the human niche may have a of who compared the ageing ovary different, i.e. slower rate of ageing than the mouse niche.
and the ageing testis with ageing somatic tissues of mice, The net outcome of a slower rate of ageing over a longer finding that ageing of germ cells generally shows a period of time (decades) is difficult to predict.
Reproduction (2014) 148 55–72 C Schwarzer, M Siatkowski and others Taken together, our data indicate that ageing of oocytes Akan P, Alexeyenko A, Costea PI, Hedberg L, Solnestam BW, Lundin S, on the protein level generally shows a different pattern of Hallman J, Lundberg E, Uhlen M & Lundeberg J 2012 Comprehensiveanalysis of the genome transcriptome and proteome landscapes of three gene expression changes than ageing on the mRNA level, tumor cell lines. Genome Medicine 4 86. () and the former does not resemble somatic ageing. The Alexa A, Rahnenfuhrer J & Lengauer T 2006 Improved scoring of functional possibilities of modern proteomics are far from exhausted groups from gene expression data by decorrelating GO graph structure.
(e.g. extended protein coverage, detection of post- Bioinformatics 22 1600–1607. ) Bouniol-Baly C, Hamraoui L, Guibert J, Beaujean N, Szollosi MS & translational modifications) and future work will contri- Debey P 1999 Differential transcriptional activity associated with bute to a sharper definition of the maternal age effect in chromatin configuration in fully grown mouse germinal vesicle oocytes.
oocytes, including, but not limited to, the correspondence Biology of Reproduction 60 580–587. ( between protein immunodetection and LC-MS/MS results.
Cao S, Guo X, Zhou Z & Sha J 2012 Comparative proteomic analysis of proteins involved in oocyte meiotic maturation in mice. MolecularReproduction and Development 79 413–422. ( Cavaleri FM, Balbach ST, Gentile L, Jauch A, Bohm-Steuer B, Han YM, Supplementary data Scholer HR & Boiani M 2008 Subsets of cloned mouse embryos and theirnon-random relationship to development and nuclear reprogramming.
This is linked to the online version of the paper at Mechanisms of Development 125 153–166. Chandrakanthan V, Li A, Chami O & O'Neill C 2006 Effects of in vitro fertilization and embryo culture on TRP53 and Bax expression in B6 Declaration of interest mouse embryos. Reproductive Biology and Endocrinology 4 61. ( The authors declare that there is no conflict of interest that Demetrius L 2005 Of mice and men. When it comes to studying ageing and could be perceived as prejudicing the impartiality of the the means to slow it down, mice are not just small humans. EMBOReports 6 S39–S44. ) research reported.
Eppig JJ & Wigglesworth K 2000 Development of mouse and rat oocytes in chimeric reaggregated ovaries after interspecific exchange of somaticand germ cell components. Biology of Reproduction 63 1014–1023.
Esteves TC, Balbach ST, Pfeiffer MJ, Arauzo-Bravo MJ, Klein DC, Sinn M & This study was supported by the priority programme Boiani M 2011 Somatic cell nuclear reprogramming of mouse oocytes (Schwerpunktprogramm) no. 1356 of the Deutsche Forschungs- endures beyond reproductive decline. Aging Cell 10 80–95. ( gemeinschaft (DFG grants BO2540/3-2 and FUE-583/2-2), by the Max Planck Society and by the BMBF Verbundprojekt ROSAge, Geiger T, Wisniewski JR, Cox J, Zanivan S, Kruger M, Ishihama Y & Mann M 2011 Use of stable isotope labeling by amino acids in cell culture as a FKZ 0315892A.
spike-in standard in quantitative proteomics. Nature Protocols6 147–157. () Gentleman RC, Carey VJ, Bates DM, Bolstad B, Dettling M, Dudoit S, Author contribution statement Ellis B, Gautier L, Ge Y, Gentry J et al. 2004 Bioconductor: open softwaredevelopment for computational biology and bioinformatics. Genome C Schwarzer performed isolation and purification of RNA from oocytes, immunoconfocal embryo imaging and participated in Grondahl ML, Yding Andersen C, Bogstad J, Nielsen FC, Meinertz H & data analysis; B Wang helped with the immunoconfocal embryo Borup R 2010 Gene expression profiles of single human mature oocytes imaging; M Siatkowski and G Fuellen performed bioinformatics in relation to age. Human Reproduction 25 957–968. data analyses; H C A Drexler designed and performed Hall VJ, Compton D, Stojkovic P, Nesbitt M, Herbert M, Murdoch A & proteomics measurements; N Baeumer performed RNA proces- Stojkovic M 2007 Developmental competence of human in vitro aged sing and microarray measurements; M J Pfeiffer helped with the oocytes as host cells for nuclear transfer. Human Reproduction 22 52–62.
mouse work; M Boiani performed oocyte collection for Hamatani T, Falco G, Carter MG, Akutsu H, Stagg CA, Sharov AA, proteomic analysis, karyotype analysis, micromanipulations, Dudekula DB, VanBuren V & Ko MS 2004 Age-associated alteration of embryo production and participated in data analysis; and gene expression patterns in mouse oocytes. Human Molecular Genetics C Schwarzer, M Siatkowski, M J Pfeiffer, G Fuellen and M Boiani 13 2263–2278. ( wrote the paper. All authors read and approved the paper.
Hodges CA, Revenkova E, Jessberger R, Hassold TJ & Hunt PA 2005 SMC1b-deficient female mice provide evidence that cohesins are amissing link in age-related nondisjunction. Nature Genetics 371351–1355. ( Hu T, Liu S, Breiter DR, Wang F, Tang Y & Sun S 2008 Octamer 4 small The authors thank Annalen Nolte for help with F9 cell culture interfering RNA results in cancer stem cell-like cell apoptosis. CancerResearch 68 6533–6540. and Amy Pavlak for editorial assistance. Excellent animal house Jiao GZ, Cao XY, Cui W, Lian HY, Miao YL, Wu XF, Han D & Tan JH 2013 support from Ludger Recker and Dr Alexandra Bu¨hler is Developmental potential of prepubertal mouse oocytes is compromised due mainly to their impaired synthesis of glutathione. PLoS ONE 8e58018. Kersey PJ, Duarte J, Williams A, Karavidopoulou Y, Birney E & Apweiler R 2004 The International Protein Index: an integrated database for proteomics experiments. Proteomics 4 1985–1988. ( Ackermann M & Strimmer K 2009 A general modular framework for gene Kocabas AM, Crosby J, Ross PJ, Otu HH, Beyhan Z, Can H, Tam WL, set enrichment analysis. BMC Bioinformatics 10 47. Rosa GJ, Halgren RG, Lim B et al. 2006 The transcriptome of human oocytes. PNAS 103 14027–14032. ( Reproduction (2014) 148 55–72 SILAC proteome analysis of mouse oocyte ageing Kosubek A, Klein-Hitpass L, Rademacher K, Horsthemke B & Ryffel GU Schwanhausser B, Busse D, Li N, Dittmar G, Schuchhardt J, Wolf J, Chen W 2010 Aging of Xenopus tropicalis eggs leads to deadenylation of a & Selbach M 2011 Global quantification of mammalian gene expression specific set of maternal mRNAs and loss of developmental potential.
control. Nature 473 337–342. ( PLoS ONE 5 e13532. Schwarzer C, Esteves TC, Arauzo-Bravo MJ, Le Gac S, Nordhoff V, Schlatt S Li L, Zheng P & Dean J 2010 Maternal control of early mouse development.
& Boiani M 2012 ART culture conditions change the probability of Development 137 859–870. ) mouse embryo gestation through defined cellular and molecular Lopes FL, Fortier AL, Darricarrere N, Chan D, Arnold DR & Trasler JM responses. Human Reproduction 27 2627–2640. ( 2009 Reproductive and epigenetic outcomes associated with aging mouse oocytes. Human Molecular Genetics 18 2032–2044. Selesniemi K, Lee HJ, Muhlhauser A & Tilly JL 2011 Prevention of maternal aging-associated oocyte aneuploidy and meiotic spindle defects in mice Lord T & Aitken RJ 2013 Oxidative stress and ageing of the post-ovulatory by dietary and genetic strategies. PNAS 108 12319–12324. ( oocyte. Reproduction 146 R217–R227. ( Ma P & Schultz RM 2008 Histone deacetylase 1 (HDAC1) regulates histone Sharov AA, Piao Y, Matoba R, Dudekula DB, Qian Y, VanBuren V, Falco G, acetylation, development, and gene expression in preimplantation Martin PR, Stagg CA, Bassey UC et al. 2003 Transcriptome analysis of mouse embryos. Developmental Biology 319 110–120. ( mouse stem cells and early embryos. PLoS Biology 1 E74. ( Ma M, Guo X, Wang F, Zhao C, Liu Z, Shi Z, Wang Y, Zhang P, Zhang K, Sharov AA, Falco G, Piao Y, Poosala S, Becker KG, Zonderman AB, Wang N et al. 2008 Protein expression profile of the mouse metaphase-II Longo DL, Schlessinger D & Ko M 2008 Effects of aging and calorie oocyte. Journal of Proteome Research 7 4821–4830. restriction on the global gene expression profiles of mouse testis and ovary. BMC Biology 6 24. Mamo S, Gal AB, Bodo S & Dinnyes A 2007 Quantitative evaluation and Steuerwald NM, Bermudez MG, Wells D, Munne S & Cohen J 2007 selection of reference genes in mouse oocytes and embryos cultured Maternal age-related differential global expression profiles observed in vivo and in vitro. BMC Developmental Biology 7 14. in human oocytes. Reproductive Biomedicine Online 14 700–708.
McGuinness BE, Anger M, Kouznetsova A, Gil-Bernabe AM, Helmhart W, Takahashi K & Yamanaka S 2006 Induction of pluripotent stem cells from Kudo NR, Wuensche A, Taylor S, Hoog C, Novak B et al. 2009 mouse embryonic and adult fibroblast cultures by defined factors. Cell Regulation of APC/C activity in oocytes by a Bub1-dependent spindle assembly checkpoint. Current Biology 19 369–380. ( Tarin JJ, Gomez-Piquer V, Pertusa JF, Hermenegildo C & Cano A 2004 Association of female aging with decreased parthenogenetic activation, Miao YL, Kikuchi K, Sun QY & Schatten H 2009 Oocyte aging: cellular and raised MPF, and MAPKs activities and reduced levels of glutathione molecular changes, developmental potential and reversal possibility.
S-transferases activity and thiols in mouse oocytes. Molecular Reproduc- Human Reproduction Update 15 573–585. ( tion and Development 69 402–410. Tatone C, Carbone MC, Gallo R, Delle Monache S, Di Cola M, Alesse E & Monti M, Zanoni M, Calligaro A, Ko MS, Mauri P & Redi CA 2013 Amicarelli F 2006 Age-associated changes in mouse oocytes during Developmental arrest and mouse antral not-surrounded nucleolus oocytes.
postovulatory in vitro culture: possible role for meiotic kinases and Biology of Reproduction 88 2. ( survival factor BCL2. Biology of Reproduction 74 395–402. () Monetti M, Nagaraj N, Sharma K & Mann M 2011 Large-scale phosphosite Tong ZB, Gold L, Pfeifer KE, Dorward H, Lee E, Bondy CA, Dean J & quantification in tissues by a spike-in SILAC method. Nature Methods Nelson LM 2000 Mater, a maternal effect gene required for early embryonic development in mice. Nature Genetics 26 267–268. Myers DD 1978 Review of disease patterns and life span in aging mice: Velde ERT & Pearson PL 2002 The variability of female reproductive ageing.
genetic and environmental interactions. Birth Defects Original Article Human Reproduction Update 8 141–154. Series 14 41–53.
Nelson JF, Gosden RG & Felicio LS 1985 Effect of dietary restriction on Vitale AM, Calvert ME, Mallavarapu M, Yurttas P, Perlin J, Herr J & Coonrod S estrous cyclicity and follicular reserves in aging C57BL/6J mice. Biology 2007 Proteomic profiling of murine oocyte maturation. Molecular of Reproduction 32 515–522. ( Reproduction and Development 74 608–616. ( Oliveri RS, Kalisz M, Schjerling CK, Andersen CY, Borup R & Byskov AG Vizcaino JA, Cote RG, Csordas A, Dianes JA, Fabregat A, Foster JM, Griss J, 2007 Evaluation in mammalian oocytes of gene transcripts linked to Alpi E, Birim M, Contell J et al. 2013 The PRoteomics IDEntifications epigenetic reprogramming. Reproduction 134 549–558. (PRIDE) database and associated tools: status in. Nucleic Acids Research 41 D1063–D1069. () Pan H, Ma P, Zhu W & Schultz RM 2008 Age-associated increase in Vogel C & Marcotte EM 2012 Insights into the regulation of protein aneuploidy and changes in gene expression in mouse eggs.
abundance from proteomic and transcriptomic analyses. Nature Developmental Biology 316 397–407. Reviews. Genetics 13 227–232. ( Pfeiffer MJ, Balbach ST, Esteves TC, Crosetto N & Boiani M 2010 Enhancing Wallace DC & Chalkia D 2013 Mitochondrial DNA genetics and the somatic nuclear reprogramming by Oct4 gain-of-function in cloned heteroplasmy conundrum in evolution and disease. Cold Spring mouse embryos. International Journal of Developmental Biology Harbor Perspectives in Biology 5 a021220. ( Pfeiffer MJ, Siatkowski M, Paudel Y, Balbach ST, Baeumer N, Crosetto N, Walther DM & Mann M 2011 Accurate quantification of more than 4000 Drexler HC, Fuellen G & Boiani M 2011 Proteomic analysis of mouse mouse tissue proteins reveals minimal proteome changes during aging.
oocytes reveals 28 candidate factors of the "reprogrammome".
Molecular and Cellular Proteomics 10 M110 004523. ( Journal of Proteome Research 10 2140–2153. Rapti A, Trangas T, Samiotaki M, Ioannidis P, Dimitriadis E, Meristoudis C, Wang S, Kou Z, Jing Z, Zhang Y, Guo X, Dong M, Wilmut I & Gao S 2010 Veletza S & Courtis N 2010 The structure of the 50-untranslated region of Proteome of mouse oocytes at different developmental stages. PNAS 107 mammalian poly(A) polymerase-a mRNA suggests a mechanism of translational regulation. Molecular and Cellular Biochemistry 340 Wilding M, Di Matteo L & Dale B 2005 The maternal age effect: a hypothesis based on oxidative phosphorylation. Zygote 13 317–323.
R Core Team 2012 R: A Language and Environment for Statistical Computing. Vienna, Austria.
Wisniewski JR, Zougman A & Mann M 2009 Combination of FASP and Ryu BY, Orwig KE, Oatley JM, Avarbock MR & Brinster RL 2006 Effects of StageTip-based fractionation allows in-depth analysis of the hippocampal aging and niche microenvironment on spermatogonial stem cell self- membrane proteome. Journal of Proteome Research 8 5674–5678.
renewal. Stem Cells 24 1505–1511. Reproduction (2014) 148 55–72 C Schwarzer, M Siatkowski and others Wu G, Han D, Gong Y, Sebastiano V, Gentile L, Singhal N, Adachi K, Zuccotti M, Boiani M, Garagna S & Redi CA 1998 Analysis of aneuploidy rate Fischedick G, Ortmeier C, Sinn M et al. 2013 Establishment of in antral and ovulated mouse oocytes during female aging. Molecular totipotency does not depend on Oct4A. Nature Cell Biology 15 Reproduction and Development 50 305–312. ( Yu G, Li F, Qin Y, Bo X, Wu Y & Wang S 2010 GOSemSim: an R package for measuring semantic similarity among GO terms and geneproducts. Bioinformatics 26 976–978. ( Received 4 March 2014 First decision 19 March 2014 Zhang P, Ni X, Guo Y, Guo X, Wang Y, Zhou Z, Huo R & Sha J 2009 Proteomic- based identification of maternal proteins in mature mouse oocytes. BMC Revised manuscript received 26 March 2014 Genomics 10 348. ( Accepted 31 March 2014 Reproduction (2014) 148 55–72

  • Outline placeholder


    Microsoft word - p349-368 _conferences abstracts_

    Proceedings of the 27th Annual Meeting of the Brazilian Embryo Technology Society (SBTE), August 29th to September 1st, 2013, Praia do Forte, BA, Brazil. Conference abstracts. Use of bovine sex sorted sperm on timed artificial insemination, in vivo and in vitro embryo production programs P.S. Baruselli1, J.G. Soares1, J.N.S. Sales2, G.A. Crepaldi1, A.H. Souza1, K.A.L. Neves1, C.M. Martins1,

    Material Safety Data Sheet 1.Product and company identification a) Product Name: (to indicate the same name or code as shown in label) Lithium Cobalt Oxide b) Recommended use of the chemical and restrictions on use: Suitable for Cathode Material of Li-ion Battery c) Manufacturer/Supplier/Distributor Information: Name, Address, Responsible department Company Name: MTI Corporation Address: 860 South 19th Street, Richmond, CA 94804, USA Tel No. : (510)525-3070