The Ashbya gossypiiEF-1α promoter of the ubiquitously used MX cassettes is toxic to Saccharomyces cerevisiae

FEBS Letters 585 (2011) 3907–3913

journal homepage: www.FEBSLetters.org

The Ashbya gossypii EF-1a promoter of the ubiquitously used MX cassettes is toxic to Saccharomyces cerevisiae Roja Babazadeh, Soode Moghadas Jafari, Martin Zackrisson, Anders Blomberg, Stefan Hohmann, Jonas Warringer ⇑,1, Marcus Krantz ⇑,1 Department of Cell and Molecular Biology, University of Gothenburg, Box 462, S-40530 Gothenburg, Sweden

a r t i c l e

i n f o

Article history: Received 30 March 2011 Revised 12 October 2011 Accepted 14 October 2011 Available online 22 October 2011 Edited by Francesc Posas Keywords: Histidine toxicity Protein overexpression HIS3 MX toxicity TAP-tag toxicity

a b s t r a c t Protein overexpression based on introduction of multiple gene copies is well established. To improve purification or quantification, proteins are typically fused to peptide tags. In Saccharomyces cerevisiae, this has been hampered by multicopy toxicity of the TAP and GFP cassettes used in the global strain collections. Here, we show that this effect is due to the EF-1a promoter in the HIS3MX marker cassette rather than the tags per se. This promoter is frequently used in heterologous marker cassettes, including HIS3MX, KanMX, NatMX, PatMX and HphMX. Toxicity could be eliminated by promoter replacement or exclusion of the marker cassette. To our knowledge, this is the first report of toxicity caused by introduction of a heterologous promoter alone. Ó 2011 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved.

1. Introduction Artificial manipulation of protein production in order to either observe the resulting change in phenotypic trait(s) [1] or to harvest the produced protein from its host [2] are fundamental approaches in modern genetics and biochemistry. Reverse genetics has proven itself tremendously successful in assigning functions to individual gene products [3,4], defining hierarchy in functional networks [5,6] and for high yield heterologous protein expression for biochemical or therapeutic purposes [7,8]. Protein overexpression is typically performed either by replacing the native promoter with a stronger one [6] or by placing the gene on a multi-copy plasmid [9]. The latter approach can be used with intact 50 and 30 UTRs to preserve native regulation and allows comparisons of tolerance to relative increase in protein abundance across all expressed proteins. The introduction of a partially defect auxotrophic marker, such as leu2-d, allows control of plasmid copy number and thus overexpression level via changes in growth

⇑ Corresponding authors. Current address: Theoretical Biophysics, HumboldtUniversität zu Berlin, Invalidenstr. 42, 10115 Berlin, Germany (M. Krantz). E-mail addresses: [email protected] (J. Warringer), Marcus. [email protected] (M.Krantz). 1 These authors equally contributed to this work.

media composition; in the case of leu2-d via the level of external leucine [10]. Complete absence of leucine forces around 100 copies of the plasmid, however, when the gene product is toxic, such expression levels are not tolerated and fitness is reduced as a consequence of the resulting ‘‘genetic Tug-of-War’’ (gToW) [11,12]. Both protein quantification and purification of the overexpressed proteins are facilitated by fusion to a translated tag sequence, e.g., Tandem Affinity Purification (TAP) or Green Fluorescent Protein (GFP) tags, for which high performance antibodies and straight forward purification protocols exist [13,14]. Unfortunately, tags from the global Saccharomyces cerevisiae collections have been reported to be toxic when introduced in a multi-copy context [12]. Using a gToW system based on leu2-d, we show that this toxicity effect constitutes a general phenomenon. However, we show that the toxicity is independent of the actual tags and due to the heterologous HIS3MX marker cassette used for tag introduction selection. Remarkably, the toxicity did not stem from the heterologous expression of HIS3 (Schizosaccharomyces pombe his5+), but from the multi-copy introduction of the promoter driving this gene; the EF-1a promoter from Ashbya gossypii. This promoter is included in many ubiquitously used heterologous markers in yeast, including the HIS3MX, KanMX, NatMX, PatMX and HphMX, giving our observations a general urgency. In addition, our findings open up for the introduction of non-toxic tagged protein constructs in a yeast multi-copy system, allowing large scale quantification of changes in trait(s) as a function of protein

0014-5793/$36.00 Ó 2011 Federation of European Biochemical Societies. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.febslet.2011.10.029

3908

R. Babazadeh et al. / FEBS Letters 585 (2011) 3907–3913

amount. Finally, they reveal a previously unknown cause of fitness loss accompanying heterologous protein expression; heterologous multi-copy promoter toxicity in the host. This introduces a new challenge and possibility in the optimisation of protein production systems.

3. Results and discussion 3.1. Multiple copies of genes tagged with the TAP or GFP cassettes are toxic Protein tagging in plasmid based overexpression systems has been reported to be toxic in S. cerevisiae [12]. To extend and more stringently examine this observation, we overexpressed 41 TAP tagged proteins involved in cellular signal transduction from the global TAP collection [13] using the gToW approach. The standard TAP and GFP tags are accompanied by a terminator sequence derived from S. cerevisiae ADH1 (TADH1) and the HIS3MX marker cassette, which consists of the HIS3 orthologue from S. pombe (his5+; henceforth SpHIS3) expressed behind the A. gossypii EF-1a promoter (henceforth AgPTEF1a) (Fig. 1A) [16]. Growth analyses on agar plates showed that overexpression of most non-tagged proteins conferred no or marginal toxic effects, in agreement with earlier observations that there is no general fitness trend from

2. Materials and methods Plasmid, strains and growth procedures have been described elsewhere [11–15] and can be found in the Supplemental materials. De novo strain construction was done as described [11] but using the GFP and TAP tagged collections as PCR templates, when appropriate. The hybrid insert containing the S. pombe his5+ flanked by the flanking regions of S. cerevisiae was synthesised. Tables S1 and S2 summarise all primers and tagged strains used in this study, respectively. DNA extraction and rtPCR used standard protocols and the manufacturer’s recommendations, see Supplements for details.

A

C

Wild Type

Empty plasmid gToW gToW-TAP

TAP-tagged

Density (OD)

8

GFP-tagged 5’FR

GFP

ORF

TADH1

3’FR

AgPTEF1a

TAP

SpHIS3

2 0.5 0.12 0

72

Time (h)

B

SC+leu

D

SC-leu

Density (OD)

Empty plasmid PDE2-gToW PDE2-gToW-TAP PDE2-gToW-GFP

Efficiency = Δ OD Lag = Δt

Rate = Δ OD/Δ t

E

Growth TAP tagged (LSCRATE)

Time (h)

0.5

0.5

Rate

Lag

Efficiency

0

-0.5

-0.5 -2 -1.5

-1.5 -4

-2.5 -3.5 -3.5

-2.5

-6 -2.5

-1.5

-0.5

0.5

-6

-4

-2

0

-3.5 -3.5

-2.5

-1.5

-0.5

0.5

Growth native (LSCRATE) LSCRATE=log2(RateEmpty Plasmid/RategToW)

Fig. 1. Overexpression of TAP or GFP tagged proteins using a multi-copy plasmid is toxic. (A) The global collections of tagged proteins contain ORFs fused with GFP or TAP tags, followed by the ADH1 terminator (TADH1) and the HIS3MX marker cassette containing the S. pombe HIS3 orthologue (SpHIS3) expressed from the EF-1a promoter of A. gossypii (AgPTEF1a). (B) Drop tests of gToW overexpressed PDE2, with and without TAP and GFP tags. The gToW plasmid is present in low plasmid copy numbers in SC +Leu medium but in high copy numbers in SC-Leu medium. The tagged but not the native constructs are toxic when present in high copy numbers. (C) Micro-cultivation growth curves for the PDE2 constructs. (D) Schematic representation of the extraction of fitness variables from growth curves. The growth rate is the slope converted to population doubling time, the growth efficiency is the total change in density, and the growth lag is the intercept between the start OD and the slope. (E) Comparison between the relative growth rate, growth efficiency and growth lag effects of 22 proteins overexpressed as native and TAP tagged proteins using the gToW method. Native constructs show a range of proliferation phenotypes when overexpressed, while tagged constructs are uniformly toxic (relative growth <0) (Table S2). The average standard deviation of phenotypes was 0.11 for both native and TAP-tagged strains.

3909

R. Babazadeh et al. / FEBS Letters 585 (2011) 3907–3913

overexpression but rather that the robustness of the cellular network varies from node to node [6,11,12]. In contrast, overexpression of the TAP or GFP tagged versions conferred very strong

B Growth rate; GFP tag

A YPD; TAP tag 106 10

Rpp2B

4

Fol2 Thg1

10

2

Tfp3

1

0.2 0 -0.2 -0.4

Rpp2B Tfp3

-0.6 -0.8 -1 -1 -0.8 -0.6 -0.4 -0.2

YPD; GFP tag

C

Rpp2B Hhf2

104

102

0.2 TAP tag

SD; GFP tag Hhf2 Rpp2B

Tfp3

102 Arp5

1 -1

-0.8

-0.6

Tfp3

-0.2

1

10

0.2

0 Arp5

4

0

Growth rate; TAP tag

Tfp3

Growth rate deletion mutant

Protein expression (molecules/cell)

growth inhibition in all cases, as exemplified by the cAMP phosphodiesterase PDE2 (Fig. 1B). To quantify these toxicity effects precisely, 22 TAP tagged and the corresponding untagged

-0.4

-0.2

0

Rpp2B

-0.4

Rpl20A Prs13A

-0.6 Shm2

-0.8 -1 0.2 GFP tag

0 Tfp3

-0.2

Rpp2B

-0.4

0.2

Growth (LSCRATE) w. tag

-0.6

Arp5

-0.8 Rtg3

-1

-1 -0.8 -0.6 -0.4 -0.2

0

0.2

Growth rate w. tag

D

gToW gToW-TAP

gToW-TAP trunc + TADH1

5’FR

TADH1

ORF

AgPTEF1a

3’FR

SpHIS3

TAP

gToW-TAP trunc

Density (OD)

E

8 Empty plasmid

2

gToW gToW-TAP gToW-TAP trunc + TADH1

0.5

gToW-TAP trunc

0.125

PDE2

GPB2 0

24

48

72

0

24

48

HSP82 72

0

24

48

72

Time (h) Fig. 2. The TAP or GFP tags are not toxic to S. cerevisiae. (A) Relative growth rate of strains with a single genomic copy of a TAP (upper panel; 188 proteins; n = 2; average standard deviation 0.12) or GFP (middle and lower panels; 178 proteins; n = 2; average standard deviation 0.05) tagged protein cultivated in synthetic defined (SD) media compared to the native expression level of that protein on SD or rich (YPD) media (Table S2). (B) Comparison of the growth rates effects caused by TAP and GFP tagging for the 97 proteins that had been tagged with both. Excluding Tfp3 and Rpp2B, no correlation between the tag effects from TAP and GFP tags was observed. (C) Comparison of the growth rate effect of gene deletions to the growth rate effect of TAP (upper panel) or GFP (lower panel) tagged versions of the same proteins. With few exceptions, tagging did not lead to loss of gene function. (D) Schematic representation of gToW constructs designed to trace the potential toxicity of the TAP tag. (E) These gToW constructs were made with three ORFs; GPB2, PDE2 and HSP82 and the growth effect measure by microcultivation.

3910

R. Babazadeh et al. / FEBS Letters 585 (2011) 3907–3913

3.2. Genomic TAP or GFP tagging confers specific but not general toxicity

overexpression strains were micro-cultivated for three days in liquid culture with automated measurements of optical density every 20 min (Fig. 1C). We evaluated the growth impact by quantification of three fitness components; growth rate, lag and efficiency (Fig. 1D). Indeed, substantially stronger toxicity effects were observed for almost all TAP tagged constructs as compared to their native versions, ranging from a twofold to an almost tenfold decrease in growth rate (Fig. 1E). The effects on the other fitness parameters were more variable, ranging from almost no effect to more than tenfold inhibition by the TAP tagged constructs. Hence, overexpression of TAP and GFP tagged proteins is indeed highly toxic, using multi-copy plasmids and the TAP and GFP cassettes from the global collections.

A

Toxic effects of TAP tagging of single proteins have been reported [17], but there are no reports of a general toxicity. To test whether high expression of the TAP and GFP tags per se is toxic, we selected 269 proteins spanning the entire range of native protein expression levels. This set includes essential, partially dispensable and completely dispensable proteins (Table S2), taken from global collections where each protein had been either TAP (188 proteins) or GFP (178 proteins) tagged in the C-terminal end [13,14]. Genomic tagging of some proteins with either TAP or GFP conferred a decrease in the growth rate (Fig. 2A) and/or efficiency (data not

gToW gToW-TAP

gToW-HIS3MX

B

5’FR

TADH1

ORF

AgPTEF1a

3’FR

SpHIS3

TAP

Density (OD)

8 Empty plasmid

2

gToW gToW-TAP gToW-HIS3MX

0.5 0.125

PDE2

GPB2 0

24

48

72

0

24

48

HSP82 72

0

24

48

72

Time (h)

C

PRPP

D

ATP

8 2 PDE2-gToW

0.5 His1

PDE2-gToW-TAP

0.125

PPi

2xH2O

His4

BY4743

8

PPi

His6

2

L-glutamine

Density (OD)

His7 L-glutamate

AICAR

his1

his5

his2

his6

8 2 0.5

His3 L-glutamate

0.5 0.125

H 2O

His5

0.125

H 2O

2-oxoglutarate

8

His2 Pi

2xNAD + H2O

His4 2xNADH + 2xH+

2 0.5

L-Histidine

0.125

his4

0

24

48

his7

72 0

24

48

72

Time (h)

E

SC+leu

SC-leu

SC-leu-his

pSBI40 gToW-HIS3 Fig. 3. The HIS3MX cassette is toxic in S. cerevisiae. (A) Schematic representation of gToW constructs created to trace the toxicity of the HIS3MX cassette. (B) Again, these gToW constructs were made with three ORFs; GPB2, PDE2 and HSP82, and the growth effect measured. (C) Schematic representation of the histidine biosynthetic pathway in S. cerevisiae. (D) Deletion mutants that disrupt the histidine biosynthetic pathway do not suppress the HIS3MX toxicity. (E) Drop test of the HIS3 gToW construct. There is no toxicity from the multi-copy expression of the native S. cerevisiae HIS3.

3911

R. Babazadeh et al. / FEBS Letters 585 (2011) 3907–3913

removed the HIS3MX marker cassette restored wild type growth with no detectable loss of reproductive fitness, verifying that the TAP tag in itself is not toxic (Fig. 2E). Curiously however, removal of the ADH1 terminator conferred strong toxicity. This is not due to transcription of specific plasmid DNA, as constructs inserted in opposite directions show similar toxicity. Instead, it points towards a general toxicity of abundant unterminated transcription and emphasises the importance of proper transcription termination. In conclusion, there is no general toxicity of the TAP and GFP proteins when expressed in baker’s yeast, nor is there a general loss of function following C-terminal tagging.

shown). However, the effects were mostly marginal and much lower than for the multi-copy gToW constructs above (Fig. 1E; note difference in axis scales). The strongest observed effect was genomic TAP tagging of Fol2, which caused a 75% increase in population doubling time, followed by tagging of Tfp3 with TAP or GFP that caused a 60% and 25% increase, respectively. More importantly, the effect did not increase with the expression level of the tagged proteins, demonstrating that it is not the high expression of the tags per se that is toxic (Fig. 2A). There was no correlation between the (lack of) effects of TAP and GFP tagging beyond Tfp3 and Rpp2B (Fig. 2B). Furthermore, there was no correlation between the effect caused by tagging and deletion, showing that the tag effect only rarely comes from a loss of function (Fig. 2C). In several cases, e.g., Tfp3, tagging gave a stronger effect than deletion, consistent with that the toxic effect was due to gain of function. To confirm that the tags per se are not toxic in a multi-copy context, we compared the effect of several versions of the PDE2, GPB2 and HSP82 gToW constructs (Fig. 2D). Truncations that

A

3.3. The HIS3MX marker cassette is toxic, but the effect is independent of histidine biosynthesis The fact that the TAP and GFP tags per se are not toxic even at high expression levels and do not lead to a general loss of protein function directly points to the introduction of the HIS3MX marker

B

gTOW-HIS3MX

Density (OD)

gTOW-AgPTEF1a

8

gToW-SpHIS3 w. HIS3 3’/5’FRs

TADH1 AgP TEF1a

2

0.5 pSBI40 gToW-SpHIS3 w. HIS3 3'/5'FRs gToW-HIS3MX

0.12

gToW-AgPTEF1a

SpHIS3 HIS3 5’FR

0

36

HIS3 3’FR

72

Time (h)

Relative leucine concentration -5

-4

-3

-2

-1

0

256

R=0.95 R=0.97

0

64

-0.8

LSCRATE

-0.4

16 gToW-HIS3MX pSBI40

-1.2

4 -5

-1.6

-4

-3

-2

-1

0

F

0.5 -0.5 -1.5 -2.5 -3.5 -3.5

-2.5

-1.5

-0.5

0.5

Growth native (LSCRATE) LSCRATE=log2(Rate

Empty Plasmid

gToW

/Rate

)

Growth difference ( LSCRATE)

E

Growth TAP tagged (LSCRATE)

Relative leucine concentration

Plasmid copy number (log)

D

C

0.5 -0.5 -1.5 -2.5 -3.5 -3.5

-2.5

-1.5

-0.5

0.5

Growth native (LSCRATE) LSCRATE=log2(Rate

Empty Plasmid

/Rate

gToW

)

Fig. 4. The A. gossypii EF-1a promoter is toxic when present in ten or more copies in S. cerevisiae. (A) Schematic representation of gToW constructs created to dissect the HIS3MX cassette. (B) The growth effect of these constructs determined by microcultivation. (C) Relative reproductive rate of a WT strain containing a gToW plasmid with the HIS3MX cassette, as a function of the relative leucine concentration (log2([leucine]/max[leucine]) where max[leucine] = 100 mg/L). (D) The copy number of the HIS3MX containing gToW plasmid scales linearly with relative leucine concentrations. The plasmid copy number was determined using qPCR of LEU2 plasmid marker and normalisation to the genomic LEU3. (E and F) The toxicity of AgPTEF1a provides a minimum toxicity and explains the major part of the discrepancy between the native and TAP tagged construct toxicity. (E) displays the effect of the AgPTEF1a toxicity in comparison with the TAP tagged constructs in Fig. 1E. The dashed red line indicates the effect of AgPTEF1a alone. (F) plots the growth defect aggravation caused by the TAP tag (LSC(TAP-tagged)  LSC(native)) as a function of the growth defect of the native protein (LSC(native)). The dashed line indicates the expected effect of AgPTEF1a.

3912

R. Babazadeh et al. / FEBS Letters 585 (2011) 3907–3913

cassette on a multi-copy plasmid as the probable cause of the toxicity. HIS3MX accompanies both the TAP and GFP tag constructs, but whereas it exists only in a single chromosomal copy in the TAP and GFP genomic collections its abundance is substantially increased when placed on a multi-copy plasmid. To investigate the potential toxicity of this heterologous marker per se, we designed a HIS3MX gToW construct (AgPTEF1a-SpHIS3), and compared its effect to the growth effect of the tagged or untagged gToW constructs with PDE2, GPB2 and HSP82 (Fig. 3A). The marker cassette alone conferred very strong toxicity, approaching but not fully equalling that of the three full length tagged constructs (Fig. 3B). Together with the absence of toxicity in the constructs lacking AgPTEF1a-SpHIS3 (Fig 2E), it clearly demonstrates that the HIS3MX cassette is the major cause of toxicity. Exposure to high concentrations of histidine is toxic and has been used as base for genetic screens [18]. High expression of expression of SpHis3 could potentially lead to high internal concentrations of histidine or one of its biosynthetic intermediates. If this causes the toxicity, it should require that the essentially linear histidine biosynthetic pathway is intact (Fig. 3C). However, deletion of HIS1-HIS7 failed to suppress the toxicity of the full length tagged PDE2 gToW construct (Fig. 3D). Hence, it is clear that the toxic effect of the HIS3MX cassette is not due to an increased flux through the histidine biosynthesis pathway. Consistently, overexpression of the native S. cerevisiae HIS3 gene with the gToW method is not toxic (Fig. 3E). Hence, the HIS3MX cassette is highly toxic when present in high numbers in S. cerevisiae, but the toxic effect is not related to the histidine biosynthetic pathway. 3.4. The A. gossypii EF-1a promoter is toxic when present at ten or more copies in S. cerevisiae The absence of any link to histidine biosynthesis left two possibilities; either the toxicity of the HIS3MX stemmed from the multi-copy presence of the AgPTEF1a or from non-histidine linked effects of SpHIS3 presence and/or expression. To test these two hypotheses, we designed gToW plasmids containing either only the AgPTEF1a or only the complete SpHIS3 ORF flanked by the 50 and 30 untranslated regions from S. cerevisiae (Fig. 4A). Expression of SpHIS3 from the S. cerevisiae HIS3 promoter did not result in any significant loss of reproductive fitness (Fig 4B). In contrast, AgPTEF1a was highly toxic, comparable to the complete HIS3MX cassette. Hence, the toxicity of the HIS3MX cassette stems in its entirety from the heterologous, multi-copy introduction of the A. gossypii EF-1a promoter with no contribution from the introduction of the S. pombe his5+ gene. We determined the tolerance limit to the HIS3MX gToW plasmid by precise quantification of the loss of growth rate as a function of decreasing leucine concentration and hence increasing plasmid copy number (Fig. 4C). The toxicity of the AgPTEF1a followed a distinct sigmoid pattern with buffer capacity in the low end and toxicity saturation in the high end. Plasmid quantification showed a linear correlation between copy number and leucine concentration (Fig. 4D). Consequently, we could estimate the tolerance limit of S. cerevisiae to introduction of the A. gossypii EF-1a promoter to around ten copies. At higher copy numbers, drastic loss of fitness occurs with saturation at around 30 copies and a threefold increase in generation time. This appears to be a minimum toxicity for all constructs including the AgPTEF1a element (Fig. 4E), and this baseline toxicity explains the bulk of the growth rate impact of the tagged proteins (Fig. 4F).

4. Conclusion Our finding that the general toxic effect of the TAP and GFP tag cassettes is caused by the A. gossypii EF-1a promoter in the HIS3MX marker has three important implications: First, AgPTEF1a is widely

used with over 400 published papers using HIS3MX alone. In addition, AgPTEF1a is used in a range of popular marker cassettes, including KanMX [19], HphMX [20], NatMX [20] and PatMX [20], making the high copy toxicity of the EF-1a promoter an almost ubiquitous issue in yeast genetics. Importantly though, there is no sign of toxicity from genomic integration in single copy. Second, it opens for use of the genomic tag collections in large scale analysis in yeast multi-copy systems. The possibility to use TAP or GFP tagged proteins to link overexpression toxicity to absolute protein abundance has previously been ruled out due to this uncharacterised toxicity effect [12], but here we show that the cassette up to and including the TADH1 terminator can be used without general toxicity. This will be a significant step forward as the correlation between mRNA and protein abundance is less than perfect [21] and as precise measurements of protein toxicity are paramount to accurate quantification of the cellular tolerance limit for each individual protein [22]. Third, it reveals a previously unknown mechanism for fitness loss during heterologous protein expression; heterologous promoter toxicity in the host. It is well known that heterologous gene expression may drastically reduce the fitness of hosts ranging from bacteria to yeast [2,23]. Toxicity seems generally to be linked to protein overproduction, but transcript toxicity effects have also been observed and the current results cannot rule out such effects [24,25]. However, the longest AgPTEF1a element with uninterrupted perfect identity to the S. cerevisiae genome was 15 bp. This is substantially shorter than elements with known RNA inhibiting functions [26], meaning that the toxicity effects of AgPTEF1a are unlikely to result from such a mechanism. Interestingly, the parts of AgPTEF1a that most closely resemble sections of the S. cerevisiae genome correspond to binding sites for transcription factors, two of which are the essential Gcr1 and Rap1 (Fig. S2). This raises the possibility that promoter titration, due to binding site overexpression, causes AgPTEF1a toxicity. While promoter titration has been reported to affect transcription [27], such effects are generally not accompanied by loss of fitness and direct toxicity of non-transcribed heterologous DNA alone has, to our knowledge, not been described before. Hence, these findings add yet another aspect into the optimisation process for heterologous protein expression. Author contribution JW and MK conceived the study with assistance from AB and SH. RB and SMJ performed experiments. MZ, JW, MK analysed the data. JW and MK drafted the paper. All authors contributed to the final version of the paper. Acknowledgements The authors would like to thank Doryaneh Ahmadpour and Lars-Göran Ottosson for assistance during rtPCR, strain construction and analysis. This work was supported by the Swedish Foundation for Strategic Research through a Japan-Sweden collaborative postdoc grant with JSPS (to MK), the European Commission FP7 through project UNICELLSYS (to SH), the Swedish Research Council (to SH), Carl Trygger Foundation (to JW) and the Royal Swedish Academy of Sciences (to JW). Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.febslet.2011.10.029. References [1] Hardy, S., Legagneux, V., Audic, Y. and Paillard, L. (2010) Reverse genetics in eukaryotes. Biol. Cell 102, 561–580.

R. Babazadeh et al. / FEBS Letters 585 (2011) 3907–3913 [2] Terpe, K. (2006) Overview of bacterial expression systems for heterologous protein production: from molecular and biochemical fundamentals to commercial systems. Appl. Microbiol. Biotechnol. 72, 211–222. [3] Eisenberg, D., Marcotte, E.M., Xenarios, I. and Yeates, T.O. (2000) Protein function in the post-genomic era. Nature 405, 823–826. [4] Giaever, G., Chu, A.M., Ni, L., Connelly, C., Riles, L., Veronneau, S., Dow, S., Lucau-Danila, A., Anderson, K., Andre, B., et al. (2002) Functional profiling of the Saccharomyces cerevisiae genome. Nature 418, 387–391. [5] Costanzo, M., Baryshnikova, A., Bellay, J., Kim, Y., Spear, E.D., Sevier, C.S., Ding, H., Koh, J.L., Toufighi, K., Mostafavi, S., et al. (2010) The genetic landscape of a cell. Science 327, 425–431. [6] Sopko, R., Huang, D., Preston, N., Chua, G., Papp, B., Kafadar, K., Snyder, M., Oliver, S.G., Cyert, M., Hughes, T.R., et al. (2006) Mapping pathways and phenotypes by systematic gene overexpression. Mol. Cell 21, 319–330. [7] Jana, S. and Deb, J.K. (2005) Strategies for efficient production of heterologous proteins in Escherichia coli. Appl. Microbiol. Biotechnol. 67, 289–298. [8] Macauley-Patrick, S., Fazenda, M.L., McNeil, B. and Harvey, L.M. (2005) Heterologous protein production using the Pichia pastoris expression system. Yeast 22, 249–270. [9] Ho, C.H., Magtanong, L., Barker, S.L., Gresham, D., Nishimura, S., Natarajan, P., Koh, J.L., Porter, J., Gray, C.A., Andersen, R.J., et al. (2009) A molecular barcoded yeast ORF library enables mode-of-action analysis of bioactive compounds. Nat. Biotechnol. 27, 369–377. [10] Schneider, J.C. and Guarente, L. (1991) Vectors for expression of cloned genes in yeast: regulation, overproduction, and underproduction. Methods Enzymol. 194, 373–388. [11] Krantz, M., Ahmadpour, D., Ottosson, L.G., Warringer, J., Waltermann, C., Nordlander, B., Klipp, E., Blomberg, A., Hohmann, S. and Kitano, H. (2009) Robustness and fragility in the yeast high osmolarity glycerol (HOG) signaltransduction pathway. Mol. Syst. Biol. 5, 281. [12] Moriya, H., Shimizu-Yoshida, Y. and Kitano, H. (2006) In vivo robustness analysis of cell division cycle genes in Saccharomyces cerevisiae. PLoS Genet. 2, e111. [13] Ghaemmaghami, S., Huh, W.K., Bower, K., Howson, R.W., Belle, A., Dephoure, N., O’Shea, E.K. and Weissman, J.S. (2003) Global analysis of protein expression in yeast. Nature 425, 737–741. [14] Newman, J.R., Ghaemmaghami, S., Ihmels, J., Breslow, D.K., Noble, M., DeRisi, J.L. and Weissman, J.S. (2006) Single-cell proteomic analysis of S. cerevisiae reveals the architecture of biological noise. Nature 441, 840–846.

3913

[15] Warringer, J., Anevski, D., Liu, B. and Blomberg, A. (2008) Chemogenetic fingerprinting by analysis of cellular growth dynamics. BMC Chem. Biol 8, 3. [16] Wach, A., Brachat, A., Alberti-Segui, C., Rebischung, C. and Philippsen, P. (1997) Heterologous HIS3 marker and GFP reporter modules for PCR-targeting in Saccharomyces cerevisiae. Yeast 13, 1065–1075. [17] Halbeisen, R.E. and Gerber, A.P. (2009) Stress-dependent coordination of transcriptome and translatome in yeast. PLoS Biol. 7, e105. [18] Klasson, H., Fink, G.R. and Ljungdahl, P.O. (1999) Ssy1p and Ptr3p are plasma membrane components of a yeast system that senses extracellular amino acids. Mol. Cell. Biol. 19, 5405–5416. [19] Wach, A., Brachat, A., Pohlmann, R. and Philippsen, P. (1994) New heterologous modules for classical or PCR-based gene disruptions in Saccharomyces cerevisiae. Yeast 10, 1793–1808. [20] Goldstein, A.L. and McCusker, J.H. (1999) Three new dominant drug resistance cassettes for gene disruption in Saccharomyces cerevisiae. Yeast 15, 1541– 1553. [21] Gygi, S.P., Rochon, Y., Franza, B.R. and Aebersold, R. (1999) Correlation between protein and mRNA abundance in yeast. Mol. Cell. Biol. 19, 1720– 1730. [22] Kitano, H. (2004) Biological robustness. Nat. Rev. Genet. 5, 826–837. [23] Samuelson, J.C. (2011) Recent developments in difficult protein expression: a guide to E. coli strains, promoters, and relevant host mutations in: Heterologous Gene Expression in E. coli (Evans, T.C. and Xu, M.-Q., Eds.), Springer Science+Business Media. [24] Iost, I. and Dreyfus, M. (1995) The stability of Escherichia coli lacZ mRNA depends upon the simultaneity of its synthesis and translation. EMBO J. 14, 3252–3261. [25] Makarova, O.V., Makarov, E.M., Sousa, R. and Dreyfus, M. (1995) Transcribing of Escherichia coli genes with mutant T7 RNA polymerases: stability of lacZ mRNA inversely correlates with polymerase speed. Proc. Natl. Acad. Sci. USA 92, 12250–12254. [26] Zamore, P.D., Tuschl, T., Sharp, P.A. and Bartel, D.P. (2000) RNAi: doublestranded RNA directs the ATP-dependent cleavage of mRNA at 21 to 23 nucleotide intervals. Cell 101, 25–33. [27] Howitt, C.A., Wilinski, D., Llinas, M., Templeton, T.J., Dzikowski, R. and Deitsch, K.W. (2009) Clonally variant gene families in Plasmodium falciparum share a common activation factor. Mol. Microbiol. 73, 1171–1185.