An inventory of yeast proteins associated with nucleolar and ribosomal components

gb-2006-7-10-r98 GBJ Research An inventory of yeast proteins associated with nucleolar and ribosomal components Staub Eike eike.staub@nucleolus.net Mackowiak Sebastian sebastian.mackowiak@molgen.mpg.de Vingron Martin martin.vingron@molgen.mpg.de

Department of Computational Molecular Biology, Max Planck Institute for Molecular Genetics, Berlin, Germany

Genome Biology 1465-6906 2006 7 10 R98 http://genomebiology.com/2006/7/10/R98 17067374 10.1186/gb-2006-7-10-r98 18 5 2006 26 7 2006 26 10 2006 26 10 2006 2006 Staub et al.; licensee BioMed Central Ltd. This is an open access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Yeast nucleolar proteins

Phylogenetic profiling and gene expression analysis of yeast proteins suggests that the nucleolus probably evolved from an archaeal-type ribosome maturation machinery by recruitment of several bacterial-type and mostly eukaryote-specific factors

Abstract

Background

Although baker's yeast is a primary model organism for research on eukaryotic ribosome assembly and nucleoli, the list of its proteins that are functionally associated with nucleoli or ribosomes is still incomplete. We trained a naïve Bayesian classifier to predict novel proteins that are associated with yeast nucleoli or ribosomes based on parts lists of nucleoli in model organisms and large-scale protein interaction data sets. Phylogenetic profiling and gene expression analysis were carried out to shed light on evolutionary and regulatory aspects of nucleoli and ribosome assembly.

Results

We predict that, in addition to 439 known proteins, a further 62 yeast proteins are associated with components of the nucleolus or the ribosome. The complete set comprises a large core of archaeal-type proteins, several bacterial-type proteins, but mostly eukaryote-specific inventions. Expression of nucleolar and ribosomal genes tends to be strongly co-regulated compared to other yeast genes.

Conclusion

The number of proteins associated with nucleolar or ribosomal components in yeast is at least 14% higher than known before. The nucleolus probably evolved from an archaeal-type ribosome maturation machinery by recruitment of several bacterial-type and mostly eukaryote-specific factors. Not only expression of ribosomal protein genes, but also expression of genes encoding the 90S processosome, are strongly co-regulated and both regulatory programs are distinct from each other.

Bioinformatics Evolution Cell biology

Background

In prokaryotes, heat and distinct ionic conditions are sufficient to assemble a ribosome from its building blocks in vitro 1. In comparison, the biosynthesis of eukaryotic ribosomes is a complicated procedure. Eukaryotic ribosomes are made in the nucleolus, the ribosome factory of a eukaroytic cell. The nucleolus is a dense compartment in the nucleus of eukaryotes where freshly transcribed ribosomal RNA (rRNA) and ribosomal proteins imported from the cytosol meet complex machinery for ribosome maturation and assembly. Ribosomal subunits leave the nucleolus in a state in which the majority of their building blocks are already incorporated 23.

Several lines of evidence suggest that ribosome biosynthesis is not the sole function of nucleoli. They have been linked to cell growth control, sequestering of regulatory molecules (for example, of the cell cycle), modification of small RNAs, mitotic spindle positioning, assembly of non-ribosomal ribonucleoprotein (RNP) particles, nuclear export, and DNA repair 2456. The wide range of different functions linked to the nucleolus is not surprising when considering the prominent position of ribosome biosynthesis with respect to cellular economy 7. It seems as if the regulation of a broad range of cellular mechanisms related to cell growth and division is linked to the ribosome biosynthesis machinery through nucleoli. The full range of molecules involved in this cross-talk is only beginning to emerge. Large scale proteomic analyses of nucleolar constituents 89 and a survey of the human nucleolar protein network 10 have recently provided a first global picture of the functional network of human nucleoli.

The baker's yeast Saccharomyces cerevisiae is a favorite eukaryotic model organism for ribosome-related research. However, knowledge about the set of proteins associated with ribosomes or their nucleolar precursors in yeast is fragmentary. There are currently 439 yeast proteins annotated as ribosomal, ribosome-associated, or nucleolar. Many have been identified in genome-scale protein localization studies 1112 as well as studies of narrower focus 131415161718. Such experiments usually represent only snapshots of cells in particular states. Furthermore, native protein localization might have been altered when proteins are expressed with fusion tags or as yeast two-hybrid baits or preys. Therefore, it is likely that many additional nucleolar or ribosome-associated proteins are still undiscovered. In support of this hypothesis, studies on the proteomes of human and mouse-ear cress nucleoli 891920 identified hundreds of proteins that were unknown before or have not yet been linked to the nucleolus. The lists of nucleolar proteins from these distantly related eukaryotes were only partially overlapping. Moreover, Andersen and colleagues 921 found that a large proportion of human nucleolar proteins localize to the nucleolus only transiently, which might also have rendered their discovery in yeast more difficult.

In this study, we aim to extend the fragmentary knowledge about the protein parts list of yeast nucleoli. We present a computational approach to predict novel nucleolar or ribosome biosynthesis proteins of yeast using data from orthologous nucleolar proteins and data sets on pairwise protein interactions or protein complexes. Using a naïve Bayesian classifier we predict novel proteins associated with nucleolar or ribosomal components at high estimated sensitivity and specificity. We study the evolution of these proteins using phylogenetic profiles across 84 prokaryotic and eukaryotic organisms, thereby complementing and extending earlier computational studies on the function and evolution of the nucleolus 2122. Finally, we investigate expression patterns of nucleolar and ribosome-associated genes to characterize the substructure of the nucleolar expression program.

Results and discussion

Prologue

This section is divided into three parts. In the first section, we describe a comprehensive list of yeast proteins that we predict to be associated with nucleolar or ribosomal components. Note that in the following paragraphs such proteins will be termed nucleolar or ribosomal component-associated (NRCA) proteins. NRCA proteins do not necessarily have to be associated with the ribosome or to be localized in nucleoli during their whole life cycle. Instead, it is possible that a predicted NRCA protein localizes to the nucleolus only temporarily or binds to nucleolar components outside the nucleolus. All proteins that associate with ribosomal and nucleolar components are the targets of our predictions. In this way we would like to capture all proteins that have the potential to exert important functions on nucleolar and ribosomal biology. In the second part of the study, the identified proteins are subjected to phylogenetic profiling, thereby providing insights into the evolution of the nucleolus and ribosome assembly. Finally, we characterize the gene expression program for NRCA proteins by comparison of expression patterns of diverse functionally or evolutionarily related sets of genes.

Prediction of novel nucleolar and ribosome-associated proteins

A prerequisite for comprehensive functional and evolutionary characterization of the nucleolus and the ribosomal machinery is a complete parts list of its proteins. We applied naïve Bayesian classification to extend the known list of 439 proteins associated with nucleolar and ribosomal components in yeast towards a complete inventory of such proteins. Before prediction of new factors, we performed an extensive cross-validation of our naïve Bayesian classifier to judge whether we are able to predict NRCA proteins with considerable accuracy (Figure 1). To this end, we built 1,000 training sets, performed a cross-validation and obtained 1,000 receiver operating characteristic (ROC) curves. The average area under the ROC curve (AUC) was approximately 0.98, which generally indicates a classifier of high performance. Based on cross-validation and ROC analysis on the training sets, we chose a conservative threshold of log(O_post) > 4 for the prediction of new NRCA proteins. During cross validation we predicted nucleolar proteins at a sensitivity of 50.4% and a specificity of 98.6% using this threshold, indicating that our predictions are very conservative.

Figure 1

Estimation of prediction accuracy

Estimation of prediction accuracy. The accuracy of predictions was estimated from 1,000 runs of 10-fold cross-validations using 1,000 alternative training sets (see Materials and methods). The threshold/working point used for the final predictions of new nucleolar proteins is marked in each plot. (a) The sensitivity (SE = TP/(TP + FN)) of our classifier is plotted over different thresholds of classifier scores (log posterior odds ratios) applied to each cross-validation run. The logarithmic posterior odds ratios indicate how likely it is under the naïve Bayesian model that a protein is an NRCA protein (positive scores) versus that it is not an NRCA protein (negative scores). A single point on the line and its error bar stems from calculations of the average sensitivity and its standard deviation obtained from 1,000 cross-validation runs using a distinct classification score threshold. Confidence intervals are ± 2-fold standard deviation intervals around the mean. Note that at the threshold that was finally used for prediction (0.4) we expect to reach a sensitivity of 50.4%. This means that we have probably still missed as many NRCA proteins as we have predicted (62). (b) The specificity (SP = TN/(TN + FP)) of our classifier is plotted over different thresholds of classifier thresholds (log posterior odds ratios) that were applied on results of each of 1,000 cross-validation runs. Confidence intervals are ± 2-fold standard deviation intervals around the mean. Note that at the finally used threshold of 0.4 the specificity reaches 0.986, meaning that we expect only 1.4% of false positives among our predictions. (c) The ROC curve of our classifier is plotted as sensitivity versus (1-specificity). Each individual data point reflects predictions at a single cross-validation run when a single prediction threshold is applied. The central line is based on averaged SE/SP values for each threshold applied. The ROC curve gives an impression of the quality of a classifier. It is a general indicator of classification performance. The bigger the AUC, the better the classifier. We obtained an AUC value of 0.98, which generally indicates a classification of high quality. The ROC curve was also the basis for the selection of our final classifier threshold, as it illustrates the trade-off between sensitivity and specificity. We chose to be very conservative (high specificity) for the sake of missing true NRCA proteins (lower sensitivity).

Out of 6,281 proteins that were not annotated as NRCA proteins before, we predicted a further 62 to be linked to nucleolus/ribosome biology (Table 1, Figure 2). The experimental evidence underlying our predictions can be encoded in a 7-bit binary data string. All data strings that occurred in our analysis are summarized in Table 2 along with the prediction results obtained for them. When sensitivity/specificity estimates of the cross-validation runs hold, we estimate that there is approximately 1 false positive prediction among the 62 proteins and that we missed about another 62 proteins by our approach. We conclude that the complete inventory of nucleolar and ribosome-associated proteins in yeast comprises 439 previously known proteins, 62 predicted in this analysis, and about another 62 proteins that remain to be discovered. Thus, we hypothesize that, in total, approximately 560 genes (more than 8% of the total gene content) encode proteins related to nucleolar or ribosomal biology in yeast.

Table 1

Classification results and annotation for 62 novel predicted nucleolar/ribosome-associated proteins

Gene

ORF

log(O)

Description

SUA7

YPR086W

0.665

TFIIB subunit (transcription initiation factor) factor E

HTA1

YDR225W

0.612

Histone H2A

HSC82

YMR186W

0.697

Heat shock protein

TIF1

YKR059W

0.699

Translation initiation factor 4A

PRP4

YPR178W

0.703

U4/U6 snRNP 52 kDa protein

KAR2

YJL034W

0.684

Component of ER translocon

HTA2

YBL003C

0.724

Histone H2A.2

AAC3

YBR085W

0.686

Mitochondrial ADP/ATP carrier - member of the mitochondrial carrier (MCF) family

RFC2

YJR068W

0.686

DNA replication factor C 41 kDa subunit

TEF1

YPR080W

0.686

Translation elongation factor eEF1 alpha-A chain cytosolic

SMX2

YFL017W-A

0.696

snRNP G protein (the homologue of the human Sm-G)

BCP1

YDR361C

0.686

Similarity to hypothetical protein S. pombe

LEA1

YPL213W

0.704

U2 A snRNP protein

HSP82

YPL240C

0.686

Heat shock protein

SMD3

YLR147C

0.699

Spliceosomal snRNA-associated Sm core protein required for pre-mRNA splicing

TIF2

YJL138C

0.686

Translation initiation factor eIF4A

None

YBR025C

0.610

Strong similarity to Ylf1p

SPT16

YGL207W

0.705

General chromatin factor

SUI2

YJR007W

0.720

Translation initiation factor eIF2 alpha chain

HSH49

YOR319W

0.702

Essential yeast splicing factor

DED1

YOR204W

0.716

ATP-dependent RNA helicase

HTB1

YDR224C

0.709

Histone H2B

HRR25*

YPL204W

0.718

Casein kinase I Ser/Thr/Tyr protein kinase

SSA2

YLL024C

0.686

Heat shock protein of HSP70 family cytosolic

SRP1

YNL189W

0.696

Karyopherin-alpha or importin

SUB2

YDL084W

0.686

Probably involved in pre-mRNA splicing

CKA1

YIL035C

0.698

Casein kinase II catalytic alpha chain

PRP43*

YGL120C

0.695

Involved in spliceosome disassembly

SUI3

YPL237W

0.721

Translation initiation factor eIF2 beta subunit

DST1

YGL043W

0.692

TFIIS (transcription elongation factor)

PRP8

YHR165C

0.721

U5 snRNP protein pre-mRNA splicing factor

PRP9

YDL030W

0.667

Pre-mRNA splicing factor (snRNA-associated protein)

SUP45

YBR143C

0.704

Translational release factor

ASC1

YMR116C

0.698

40S small subunit ribosomal protein

DBP2*

YNL112W

0.719

ATP-dependent RNA helicase of DEAD box family

CKB2

YOR039W

0.710

Casein kinase II beta chain

YRA1

YDR381W

0.720

RNA annealing protein

GCD11

YER025W

0.609

Translation initiation factor eIF2 gamma chain

TFG2

YGR005C

0.695

TFIIF subunit (transcription initiation factor) 54 kDa

TOP1*

YOL006C

0.693

DNA topoisomerase I

BRR2

YER172C

0.708

RNA helicase-related protein

RVB1

YDR190C

0.709

RUVB-like protein

MLP1

YKR095W

0.686

Myosin-like protein related to Uso1p

HTZ1

YOL012C

0.685

Evolutionarily conserved member of the histone H2A F/Z family of histone variants

ATP2

YJR121W

0.685

F1F0-ATPase complex F1 beta subunit

SMD2

YLR275W

0.688

U1 snRNP protein of the Sm class

PRP3

YDR473C

0.704

Essential splicing factor

EFT1

YOR133W

0.682

Translation elongation factor eEF2

HTB2

YBL002W

0.690

Histone H2B.2

TEF4

YKL081W

0.718

Translation elongation factor eEF1 gamma chain

HHF2

YNL030W

0.695

Histone H4

Predictions based solely on protein interactions

RPO21

YDL140C

0.728

DNA-directed RNA polymerase II 215 kDa subunit

DHH1

YDL160C

0.714

Putative RNA helicase of the DEAD box family

CFT1

YDR301W

0.731

Pre-mRNA 3-end processing factor CF II

KAP95

YLR347C

0.689

Karyopherin-beta

SPT5

YML010W

0.732

Transcription elongation protein

TAF14

YPL129W

0.733

TFIIF subunit (transcription initiation factor) 30 kDa

RPB3

YIL021W

0.728

DNA-directed RNA-polymerase II 45 kDa

RPO31

YOR116C

0.729

DNA-directed RNA polymerase III 160 kDa subunit

TIF4631

YGR162W

0.734

mRNA cap-binding protein (eIF4F) 150K subunit

PRP24

YMR268C

0.734

Pre-mRNA splicing factor

RET1

YOR207C

0.731

DNA-directed RNA polymerase III 130 kDa subunit

The data used for classification and the detailed prediction results are listed for all 62 proteins that passed our threshold of O_post> 0.4. These proteins had not been annotated as associated with nucleolar or ribosomal components before, but were classified as such in our analysis. A literature survey for the predicted proteins revealed that for four proteins a role in the nucleolus and ribosome biogenesis had already been established (see Note added in proof). The lower part of the table lists 11 proteins that were predicted as NRCA proteins solely on the basis of shared participation in complexes or interactions. For these proteins, we do not necessarily predict a nucleolar localization, but direct interaction with nucleolar/ribosomal components at least under one specific cellular condition at an unspecified locus within the cell. *Four proteins for which recent articles have confirmed a role in ribosome biogenesis or the nucleolus. The results are supplemented by a concise annotation for each protein from the Comprehensive Yeast Genome Database (CYGD) [72]. The header line contains abbreviations describing the column content: Gene, gene symbol of yeast gene; ORF, yeast open reading frame ID; Hs, orthology to human nucleolar protein; At, orthology to mouse-ear cress nucleolar protein; It, link to nucleolar protein via Y2H interaction in Ito dataset; Ue, link to nucleolar protein via Y2H interaction in Uetz dataset; Ga, link to nucleolar protein via participation in a complex in Gavin data set; Ho, link to nucleolar protein via participation in a complex in Ho data set; Kr, link to nucleolar protein via participation in a complex in Krogan data set; log(O), average posterior odds ratio from all prediction runs in which the protein was not used for training; Description, concise description of protein function.

Figure 2

Phylogenetic profiling of novel nucleolar/ribosome-associated proteins

Phylogenetic profiling of novel nucleolar/ribosome-associated proteins. Phylogenetic profiles of 62 previously unrecovered nucleolar/ribosome-associated proteins of yeast across 84 organisms. The profiles were generated using the best reciprocal hit method with yeast as a reference organism (see Materials and methods). Abbreviations given on the top of the plot represent organism names (first three letters for genus and first three letters of species names; see Materials and methods for a translation of abbreviations into organism names). Further taxonomic annotation is given on the bottom of the plot. Yeast open reading frame identifiers are given on the left side, and gene names and descriptions are given on the right side of the plot. The significance of sequence similarity is visualized by different shades of gray that reflect the logarithmic expectation (E) value from reciprocal BLAST searches (shown at the bottom of the figure). Here, the E values of BLAST searches using target proteome sequences as queries versus the yeast proteome reference database are shown. The genes are ordered according to hierarchical clustering (see Materials and methods).

Table 2

Summary of effective prediction rules obtained by Bayesian classification

Prediction: associated with nucleolar or ribosomal component?

Yes