Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Gene expression response to misfolded protein as a screen for soluble recombinant protein

Gene expression response to misfolded protein as a screen for soluble recombinant protein Abstract Proper protein folding is key to producing recombinant proteins for structure determination. We have examined the effect of misfolded recombinant protein on gene expression in Escherichia coli. Comparison of expression patterns indicates a unique set of genes responding to translational misfolding. The response is in part analogous to heat shock and suggests a translational component to the regulation. We have further utilized the expression information to generate reporters responsive to protein misfolding. These reporters were used to identify properly folded recombinant proteins and to create soluble domains of insoluble proteins for structural studies. Introduction Structural genomics efforts are an important element of evaluating gene function. Protein expression and purification are key processes in these studies, and are often limited by the ability to produce properly folded recombinant protein. Escherichia coli is a common expression host that often makes misfolded protein when obliged to overproduce non-native gene products. It is easy to overlook that this workhorse is itself an evolved biological organism reacting to a stress, namely overproduction of a foreign gene product. Gene expression in response to protein misfolding is key to understanding the reaction to this stress and potential solutions to the practical aspects of recombinant protein expression. A variety of environmental stresses cause protein denaturation resulting in a classical heat-shock response. Unfolded protein is the stimulus for the change in gene expression (Parsell and Sauer, 1989). It is estimated that 10–15% of the proteins in E.coli are at least partially denatured under heat stress (Mogk et al., 1999). DnaK interacts with these proteins and initiates a refolding pathway involving DnaJ and GrpE (Beckmann et al., 1990; Hartl, 1996; Gething, 1997; Goloubinoff et al., 1999). DnaK is also a central transcription-regulator by negatively influencing the transcription factor RpoH (Liberek and Georgopoulos, 1993; McCarty et al., 1996). DnaK–RpoH interaction provides a feedback mechanism and is the proposed switch by which the cell senses the level of unfolded protein and can respond through appropriate gene expression. A major source of unfolded proteins in vivo is the nascent chain of translating genes. Gene expression in response to translational stress is related to the heat-shock response (Parsell and Sauer, 1989), but has not been closely examined. Genome arrays used to study global gene expression in E.coli at elevated temperature (Richmond et al., 1999) show changes in a large number of transcripts. To study translational misfolding rather than heat stress, we used recombinant expression of folded and misfolded proteins as a stimulus. Our results indicate a unique set of genes respond to translational misfolding. The nature of these genes implies that many known heat-shock and chaperone proteins, as well as ribosome-associated proteins suggestive of translational regulation, are induced to deal with translational stress. We have applied these results in a practical way by developing a screen to improve our ability to generate recombinant proteins for structural studies. Misfolded protein often appears as insoluble aggregates when overexpressed in E.coli. Several recent approaches have been used to detect soluble recombinant protein. They utilize protein fusion to a reporter gene such as CAT (Maxwell et al., 1999), GFP (Waldo et al., 1999), or lacZα (Wigley et al., 2001) as a measure of the amount of soluble protein expressed. Though convenient, these reporters have the potential to be biased by the nature of the fusion. The reporter portion of the fusion may alter the solubility of the target protein, either positively or negatively, giving unexpected results when expressing in the absence of the reporter fusion. By utilizing a sensory and regulatory system already existing in the host, we have created a general screen for detecting misfolded, and typically insoluble, recombinant proteins without the requirement of a direct protein fusion. We have also demonstrated the utility of this approach combined with mutagenesis to create soluble fragments of recombinant proteins in E.coli. The approach provides a general means of providing folded domains for structural studies. Materials and methods Cloning Clones expressing properly folded or misfolded human proteins were obtained from the GeneStorm collection (Invitrogen). Clones containing the Unigene accession numbers L35545, U18291, M94856, M22146, D87116, M63167, M68520, M60527, M36881, M36981, U35003, S79522, X73460, D14520, U14968, M86400 were provided in the pBADThio vector to provide arabinose-inducible expression. Thermotoga maritima genes were amplified from genomic DNA and cloned into the expression vector pMH1 which encodes a 12 amino acid N-terminal tag containing a 6X-histidine repeat for purification and detection. Reporter vectors were constructed by insertion of a PCR amplifer of 300 bp upstream of the ibpAB, ybeD, yhgI or yrfGHI genes upstream of β-galactosidase in a pACYC184 derivative. Rep68 was cloned from ATCC 68066 containing the entire genome of the human adeno-associated virus 2 (AAV2). Putative domains comprised of bases 1–646, 647–1456 and 1457–1611 were amplified from the full-length template and cloned into pMH1. The above template was also used in amplifications of the full-length gene for fragmentation. Two micrograms of the Rep68 amplifer were used in each of five fragmentation reactions containing 1, 0.1, 0.01, 0.001 or 0 U DNase I (Boeringer Mannheim) as well as Pfu polymerase and dNTPs. Reactions were set up on ice with the DNase added immediately prior to temperature cycling in an MJ Research thermocycler according to the following: 10 min at 25°C, 15 min at 95°C and 30 min at 72°C. Each reaction was run on a 1% agarose gel and fragments corresponding to 1600–1000, 1000–850, 850–600 and 600–300 bp were extracted. Each pool was used as above for blunt cloning and ligation into pMH1 as above and introduced into the reporter cell line HK 57 for screening. Thermotoga proteins used for expression studies evaluating proteins of known expression characteristics were cloned into pMH1 as described above. Coding regions were introduced for: TM0560, TM0414, TM0574, TM0703, TM0554, TM0556 (soluble expression); TM0688, TM0633, TM0712, TM0343, TM0218, TM0294 (mixed expression); TM0289, TM0564, TM0540, TM0425, TM0731, TM0413 (insoluble expression). Cell growth and protein expression Escherichia coli strains MG1655 (F−lam rph1) and KY1429 (F-araD139 Δ(argF-lac)169 lam flhD5301 fruA25 relA1 rpsL150 zhh50::Tn10 rpoH606(ts) deoC1) were transformed with expression plasmids encoding M36881 [human lymphocyte-specific protein tyrosine kinase (LCK)] or M86400 [human phospholipase A2 (PLA)] for expression profiling. Cells were cultured at 37°C in Luria broth (LB) containing ampicillin. Protein expression was induced by the addition of l-arabinose to a final concentration of 0.1% for 1 h. KY1429 cells were cultured as above except initial growth was performed at 32°C followed by a shift to 42°C for non-permissive expression of rpoH606. Top10 cells (F−mcrA Δ(mrr-hsdRMS-mcrBC) ϕ80lacZΔM15 ΔlacX74 deoR recA1 araD139 Δ(ara-leu)7697 galU galK rpsL endA1 nupG) containing the ibpAB promoter fusion (pHK57), were transformed with expression constructs listed above. β-Galactosidase assays were performed essentially as described by Miller (Sambrook et al., 1989). Fractionation of soluble and insoluble proteins was performed by centrifugation. Cultures were resuspended in 50 mM Tris pH 7.9, 50 mM NaCl, 1 mM MgCl2, 3 mM methionine and sonicated for 2 min on ice. Cell debris and insoluble protein aggregates were pelleted by centrifugation at 3000 g for 15 min. The soluble fraction was removed and the pellets resuspended in an equivalent volume of lysis buffer. Probe preparation and hybridization and analysis of labeled mRNA Labeled mRNA was prepared and hybridized to an E.coli whole genome array (Affymetrix) essentially as described previously (Lockhart et al., 1996; Wodicka et al., 1997). This gene chip contains 25-mer oligonucleotide probes for each of the 4290 known E.coli genes. Standard Affymetrix GeneChip analysis software was used to measure individual gene expression and to perform pairwise comparison of gene expression levels for pre-induction and post-induction samples. Comparisons of changes in gene expression for properly folded and misfolded genes were analyzed for individual gene probe sets. Microplate solubility screening Ninety-six-well microplates containing 200 μl of LB with 100 μg/ml ampicillin and 34 μg/ml chloramphenicol were inoculated with single colonies from above and grown overnight with shaking at 37°C. Overnight cultures were used to inoculate 200 μl of the same media and incubated at 37°C until reaching an average OD600 of 0.5. Cultures were induced with a final concentration of 0.2% arabinose. After 30 min, a cocktail of ceftriaxone and cefotaxime was added to each well to a final concentration of 10 μg/ml of each and the plates were incubated for an additional 1.5 h. These antibiotics provide a convenient method of lysis in microplate format without the need to add detergents that might solublize misfolded protein. Cultures were harvested after 2 h total of induction by centrifugation at maximum speed for 15 min to pellet cell debris on the bottom of the wells. The soluble lysate was then separated by transferring 25 μl into one set of clean microplates for β-galactosidase activity screens and 75 μl into Nunc Maxisorp ELISA plates for Ni-HRP screening. β-Galactosidase activity screening of lysates was performed using a variation of the Miller protocol (Maxwell et al., 1999). A 50 μl aliquot of 4× Z-buffer and 50 μl of 4× ONPG were added to microplates containing 25 μl of soluble lysate. After development of yellow color in positive control wells, the reaction was quenched with 75 μl of 1 M Na2CO3 pH 8. The A420, A550 and reaction times were recorded and used along with the OD600 data to calculate β-galactosidase activity (Maxwell et al., 1999). Ni-HRP screening was performed similar to an ELISA. A 75 μl aliquot of lysate plus 25 μl TBS was bound overnight at 4°C to a microtiter plate and blocked with 1% (w/v) BSA in TBS for 4 h at 25°C. Plates were then washed 3× with TBST and 100 μl of Ni-HRP conjugate (KPL Labs) was added at a dilution of 1:2500 and incubated 1 h at 25°C. The plates were then washed with TBST and 100 μl of the HRP substrate (KPL Labs) was added and color was allowed to develop until the positive control well was deep blue. The reaction was quenched with 100 μl of 1 N HCl and the A420 determined. Solubility scores were calculated by weighting the Ni-HRP A420 readings such that the experimental mean was one order of magnitude greater than the mean of the β-galactosidase activity scores, then dividing the Ni-HRP absorbance by the β-galactosidase activity. This calculation was found empirically to provide good distinction between soluble and insoluble proteins. Results Analysis of gene expression To examine gene expression as a result of misfolded protein, representative genes were cloned as fusion proteins to thioredoxin under control of the tightly regulated arabinose promoter. PLA is almost entirely soluble, as determined by cell lysis and fractionation by centrifugation (Figure 1). Further evidence of proper folding of this protein was obtained through dynamic light scattering of purified protein and the ability to crystallize it from a single affinity purification step (unpublished data). Under equivalent expression conditions, LCK is expressed almost exclusively as insoluble protein. Both proteins were expressed at sufficient levels to be the predominant translation product. mRNA preparations from induced and non-induced cultures were prepared and used to probe for gene expression. Recombinant protein expression within E.coli is predicted to cause a substantial change in gene expression. Indeed, a comparison of gene expression with a pre-induction control shows 6% of total genes with >3-fold differences in expression regardless of whether the expressed protein is soluble or insoluble. In the case of insoluble recombinant protein, 27 genes show >10-fold changes in expression, as compared to 10 genes in the case of the soluble recombinant protein. A comparison of the two profiles identifies 52 genes listed in Table I showing >3-fold changes, that are unique to the insoluble case. These genes, then, are likely responsive to misfolded protein in the cell and may play a role within E.coli in dealing with this translational stress. The heat-shock transcription factor RpoH is normally repressed by interaction with the chaperone protein DnaK. In the presence of misfolded protein, DnaK binds to that protein thereby allowing RpoH to stimulate transcription of heat-shock promoters (Liberek and Georgopoulos, 1993). Upstream regions of many of the induced genes in Table I show the presence of RpoH-dependant promoter sequences. Further evidence of the important role played by RpoH is provided by expression profiling results performed from an rpoH606 mutant (KY1429) expressing misfolded LCK protein compared to a non-expressing control. A strikingly different expression profile is seen in the case of the rpoH606 mutant (Tables I and II). The majority of the genes induced by the misfolded protein in the wild-type strain are poorly induced in the rpoH606 mutant indicating that they are directly or indirectly under the control of this transcription factor. Induction of heat-shock genes Not surprisingly, many of the genes induced by translational misfolding have known chaperone activity. These include the well characterized dnaJ, dnaK and grpE genes. The corresponding proteins interact as a complex with misfolded or denatured protein in an ATP-dependant repair process. Likewise, mopAB genes forming the GroELS folding repair complex are induced under translational misfolding conditions. IbpAB are small heat-shock polypeptides associated with inclusion body aggregates of recombinant protein (Allen et al., 1992). Whereas they do not appear to behave as folding chaperones directly, they bind misfolded protein and interact with the DnaJK GrpE proteins as a chaperone system (Thomas and Baneyx, 1998). Hsp33, the gene product of the yrfI gene was recently identified as a chaperone protein responsive to oxidizing conditions (Veinger et al., 1998). Genes implicated in degradation of denatured protein are also induced by translational misfolding. The lon, clpBP and hslUV protease genes are expressed at increased levels. Under normal cell growth these proteases serve an important recycling function. Insoluble aggregates are relatively resistant to proteolysis and this recycling pathway is ineffective for recombinant protein expression. Induction of ribosome-associated genes Other heat-shock genes associated with the ribosome are induced under conditions of translational misfolding. Hsp15 (yrfH) binds RNA (Sambrook et al., 1989) and is associated with free 50S ribosomal subunits containing a nascent polypeptide chain (Korber et al., 2000). Heat shock also increases the level of Hsp15-binding implying increased dissociation of 50S and 30S subunits. Further suggestion of ribosomal dissociation comes from the induction of ftsJ (rrmJ). The ftsJ gene product is an RNA methylase specific for 23S rRNA only when contained in the 50S ribosomal subunit (Caldas et al., 2000a,b; Puglisi et al., 2000). This enzyme methylates 23S rRNA at position 2552 located within the peptidyl transferase center of the ribosome (Caldas et al., 2000a). Mutants in ftsJ lack methylation of 23S rRNA and show up to 65% decrease in ribosomal activity corresponding to dissociation of the 50S and 30S subunits (Caldas et al., 2000b). Particularly striking in the rpoH mutant is the large increase in transcripts of the cold-shock proteins (CSPs). Table II shows the response of CSP transcripts to misfolded protein in the rpoH mutant and the wild-type rpoH strain. These genes are not affected by heat shock (9), but are associated with a transient halt of translation. CSPs are RNA-binding proteins which act as chaperones for untranslated message (Jiang et al., 1997; Wang et al., 1999) and provide anti-termination activity (Bae et al., 2000). Increased expression of CSPs under conditions that reduce chaperone expression (rpoH606) is an indication of paused translation. Taken together, these results suggest a translational regulatory response to misfolded protein. Such regulation might involve rRNA demethylation, as a consequence of translational misfolding. This hypothesis is an interesting regulatory mechanism currently under investigation. Other induced genes yccV, yhdN and yrfG have been shown to increase expression under heat-shock conditions but are of unknown function. In addition to these known heat-shock genes, yagU, yciS, ybeD, yejG and yhgI show increased expression. Most of these proteins are relatively small and generally acidic. One speculation is that some of these proteins perform a similar role to IbpAB in the direct recognition and sequestering of misfolded protein. However, only IbpAB have been associated with misfolded and aggregated protein. Induction levels of ibpAB are much higher and these other proteins may be present at lower levels. Interestingly, knockout mutations of ibpAB have relatively little affect on cell growth and viability (Thomas and Baneyx, 1998; data not shown) suggesting some functional redundancy within the cell. Genetic reporter of protein folding To confirm the profiling results and facilitate experimentation with a larger number of recombinant proteins, we cloned the promoter regions from ibpAB, ybeD, yhgI and yrfGHI into a β-galactosidase reporter vector. In each case, increased β-galactosidase activity was observed when expression of the misfolded protein LCK was induced whereas the folded protein PLA showed no increase in activity. These results were further extended using a set of eight misfolded proteins and six properly folded proteins co-expressed in the presence of the ibpAB-promoter β-galactosidase fusion. In each case, increased β-galactosidase activity corresponded to expression of misfolded protein (data not shown). A more detailed characterization is shown below. The response observed, then, appears to be a general result of protein misfolding rather than a specific response to any particular protein. These reporters provide a simple means of identifying misfolded protein through a sensitive enzymatic assay and the ibpAB promoter fusion was chosen as the reporter for further studies. ELISA-like assay for soluble protein For identifying protein derivatives that have improved folding properties in a recombinant environment, we also developed an ELISA-like assay compatible with high-throughput screening instrumentation. To evaluate soluble protein levels in a high-throughput system, non-denatured cell lysates must be prepared using conditions compatible with rapid screening in microplates. In lieu of the detergent or organic lysis, we added an antibiotic cocktail to each well to induce lysis. Soluble protein fractions were removed, bound to microtiter plates, and recombinant protein detected via binding of a Ni-HRP conjugate to a 6X-histidine N-terminal fusion. It should be noted that the His-tag may not be uniformly accessible among recombinant proteins. A negative Ni-HRP response, therefore, may not be indicative of an absence of soluble protein, but the protein fold may occlude access to the His-tag. However, we have not observed this to be a common problem. This assay, then, provides a measure of the levels of soluble recombinant protein without the need to run an SDS gel and in a form that is compatible with a HT-screen and the β-galactosidase assay. Testing proteins with pre-determined expression characteristics As part of our effort aimed at cloning, expressing and characterizing the total proteome of T.maritima, we tested the efficacy of the reporter on a set of 18 T.maritima proteins (six soluble, six insoluble and six mixed solubility). To optimize assay parameters, strains were arrayed in 96-well plates and assayed in triplicate at three induction levels (0.02, 0.2 and 2% arabinose) and at four post-induction time points for addition of the lysis-promoting antibiotics (t=0, 30, 60 and 120 min after addition of arabinose.) Figure 2 shows the averaged results for triplicate plates (soluble, insoluble and mixed) for the 0.2% arabinose induction. Both the insoluble and the mixed pools showed >4-fold higher β-galactosidase activity than the soluble pool. Conversely, the soluble pool showed a >10-fold higher response in the Ni-HRP assay opposed to the insoluble pool. The mixed pool, comprised of proteins expressed approximately equally in both soluble and insoluble fractions, showed Ni-HRP binding approximately half the intensity of the soluble pool. Although either lack of β-galactosidase or presence of Ni-HRP activity alone could be used as a measure of soluble protein, we chose a ratio of the two activities, a more effective and convenient screen. Testing proteins with unknown expression characteristics We next applied this screen to a large set of proteins with unknown folding properties. We performed the screen under the optimal conditions noted above on 186 T.maritima proteins not previously characterized for expression. The results of this screen are summarized in Table III. SDS–PAGE of eluates from nickel-chelating resin and the dissolved insoluble fractions for each clone was performed along with corresponding β-galactosidase activity, Ni-HRP response and solubility scores for 186 clones. Based on the results of the gels, 57 clones did not overexpress a visible protein band, 62 clones expressed predominantly soluble protein, 27 expressed predominantly insoluble aggregates and 46 expressed approximately equally in both soluble and insoluble fractions. A comparison of β-galactosidase activity to the Ni-HRP assay is shown in Figure 3. Points are categorized by SDS gel analysis of the soluble and insoluble protein fractions. The screen positively identified 54 of 62 (87%) soluble proteins. Seven of the eight remaining proteins that were soluble according to the gels had low Ni-HRP assays, most likely due to inaccessibility of the His-tag in these fusion proteins. Taken alone, the β-galactosidase activity measurement identified 22 of 27 (81%) insoluble proteins. Those proteins showing partial solubility showed variable solubility scores, suggesting partial folding is inducing β-galactosidase through the reporter. This assay, then, provides an effective and convenient means of classifying folding characteristics. Identification of soluble protein domains The true utility of this system lies in the ability to identify variants of full-length gene products, either mutants or domains, based on improved properties. For structural and biochemical studies, we tested the ability of this screen to identify soluble fragments of Rep68 (GI: 209617), an adeno-associated virus non-structural protein possessing various activities related to the integration of the viral genome into target DNA. This protein previously had been found to express predominantly as unfolded aggregates in E.coli. We performed both a random approach and a rational approach based on selection of domains with regard to homology. Three domains of Rep68 were selected after an RPS-BLAST search (Altschul et al., 1997) identified an internal domain possessing homology to a parvovirus non-structural protein, NP-1. This information, combined with a Kyte–Doolittle hydropathy plot (Figure 4), was used to assign the 5′ and 3′ cut-offs for each domain. The remaining N-terminal and C-terminal residues comprised the other two domains and did not possess significant homology to any other proteins in the database. Random fragments of Rep68 were also generated for screening by DNase fragmentation. The three predicted domains and randomly generated fragments were screened to identify soluble fragments of Rep68. None of the three predicted domains were identified as soluble (Figure 4). Concurrently, 564 randomly generated fragments of rep68 were also screened. One fragment returned a significantly high solubility score (Table III). This clone was verified by large-scale expression and showed expression in both the soluble and insoluble fractions. Subsequent sequencing of the identified clone verified that it was comprised of a fragment of rep68 corresponding to amino acids 1–95 (Figure 4). The identified fragment showed substantial improvement in solubility over the full-length protein and is being tested in crystallization trials. Discussion Most gene expression studies of in vivo protein folding have focused on denaturation as a result of environmental stress. This response is essential in vivo to deal with ever-changing environmental and non-ideal growth conditions. Translational folding issues are equally important to the cell since every protein as it is being translated is essentially in an unfolded state. Expression of unnatural proteins, either through recombinant means or mutation is a `stress' in itself. We show that the cellular response to translational misfolding, like heat shock, involves many known chaperone genes with a clear inference how these gene products are involved in the folding of the nascent polypeptide. In addition, other non-heat-shock genes and genes of unknown function are induced. These too may be involved in the folding process. Our results suggest both transcriptional and translational regulation. The DnaK–RpoH interaction is well characterized and appears to be the major regulator of the transcriptional response. The altered expression of genes implicated in translational stalling and ribosomal dissociation is intriguing and implies that these effects might be a result of translationally misfolded protein. These genes include yrfH which associates with the 50S ribosomal subunit (Korber et al., 2000). cspABGI are induced in rpoH606 suggesting translational stalling in the absence of induced chaperones. Also included is ftsJ, which is known to methylate the 23S rRNA of 50S subunits resulting in higher affinity of the two subunits for each other (Caldas et al., 2000a,b). Ribosome structure shows that the location of the methylation site of ftsJ, position 2552 of the 23S rRNA, is intriguingly close to the peptidyl transferase center (Puglisi et al., 2000) making it an obvious potential regulator mechanism for a ribosomal sensor of misfolded protein. Such a ribosomal sensor is not unprecedented as demonstrated by the well characterized stringent response to uncharged tRNAs during translation (Cashel et al., 1996). Ribosomal stalling provides a mechanism to allow time for chaperone synthesis and recruitment thereby preventing irreversible aggregation. In this way, the cell would retain an additional salvage pathway where the emerging protein was held in the relatively protected environment of the translating ribosome until sufficient chaperones could be recruited. The differentially regulated genes identified provide a valuable opportunity to create novel reporters of the folding state of cellular proteins as a whole and overexpressed, recombinant proteins in particular. Our reporter assay differs from others recently described by not relying on direct coupling of the reporter gene to the target, thereby limiting potential interference by the reporter. The combination of the Ni-HRP and β-galactosidase assays provides an effective means of assaying soluble recombinant proteins in a high-throughput way. We have extended this system to identify mutants and truncations of single gene products as a strategy to identify soluble domains of otherwise misfolded, aggregated proteins. Using this approach, we have identified soluble fragments of Rep68 and anticipate that this assay will provide a general means of isolating recombinant protein suitable for structure/function work. Table I. Fold change in gene expression for genes unique to misfolded response       Recombinant proteina  rpoHb  Heat shockc        None  Folded  Misfolded      aRNA preparations were made from cell lines either not expressing recombinant protein or expressing folded or misfolded protein were compared topre-induction controls to determine fold changes in transcript levels.  bFold changes in transcript levels for indicated genes in cells containing a mutant rpoH gene when overexpressing misfolded protein.  cExpression data for the corresponding heat-shock genes as described by Richmond et al. (Richmond et al., 1999).  Heat-shock genes  ibpA  Chaperone  −1.5  −1.7  74.4  14.3  297.4    ibpB  Chaperone  1.4  2.7  40.0  10.4  327.2    yrfH  Ribosome-associated HSP  1.6  −3.2  28.3  2.8  51.3    yccV  Unknown  −1.9  −2.4  19.3  3.8  34.3    fxsA  Suppresses F exclusion of phage T7  −3.8  2.1  22.3  2.0  50.7    dnaK  Chaperone  1.8  3.2  16.6  3.8  58.5    htpG  Chaperone  −3.5  −2.6  13.2  3.0  33.8    clpP  Protease  2.6  −3.6  11.8  2.7  3.3    yhdN  Unknown  1.5  3.9  11.1  3.9  9.5    clpB  Protease  2.3  −1.3  9.7  7.0  36.5    hslV  Protease  1.1  1.7  7.4  3.5  16.2    mopA  Chaperone  1.9  −1.2  6.4  1.7  37.9    lon  Protease  −1.0  1.3  6.0  2.9  20.3    mopB  Chaperone  2.4  1.1  5.8  1.7  77.5    dnaJ  Chaperone  1.2  2.7  5.6  4.2  85.3    yrfG  Unknown  −1.0  −1.3  5.0  2.1  12.1    htpX  HSP, unknown  −2.5  −1.3  4.9  2.6  36.1    hslU  Protease  −1.0  −2.1  4.7  3.0  10.3    grpE  Chaperone  −1.4  1.5  3.9  2.6  24.1    yrfI  Chaperone  −1.5  −1.1  3.6  2.7  21.6    rrmJ  rRNA methylase  −1.2  1.5  3.2  3.0  9.1  Other induced  yagU  Unknown  2.6  2.4  17.4  3.1      yciS  Unknown  1.8  2.6  14.8  5.4      ybeD  Unknown  1.3  3.8  12.0  1.8      araE  Arabinose transport  1.2  3.1  11.6  17.7      yojH  Unknown  −1.9  5.2  9.7  −3.0      yejG  Unknown  1.6  3.0  7.3  5.0      exbB  Uptake of enterchelin  −1.6  −4.3  6.4  1.7      yhgI  Unknown  1.0  −1.0  5.3  2.2      proP  Proline transport  1.0  −1.0  5  5.1      kgtP  α-Ketoglutarate permease  −1.2  2.4  4.2  3.2    Downregulated  recR  Recombination and DNA repair  −3.3  −9.0  −17.9  −3.0      lamB  Maltose uptake  12.4  −4.8  −9.9  −5.7      glpD  Glycerol-3-phosphate dehydrogenase  −2.9  −2.3  −9  −8.1      yfiD  Unknown  5.3  −3.2  −8.6  −1.4      rbsC  d-Ribose transport  3.3  −5.9  −8.2  −3.0      glpF  Glycerol facilitator  −3.9  −3.2  −8  −10.3      yqjE  Unknown function  −1  −7.6  −7.7  1.1      ftsZ  Cell division; initiation of septation  −1.4  −8.0  −7.2  1.6      ycfN  Unknown function  −1.5  −1.3  −7.1  1.3      feoA  Ferrous iron uptake  2.1  −4.3  −7.1  1.1      ybjC  Unknown function  −1.5  −5.2  −6.9  −2.4      yccA  Unknown function  −3.6  −5.0  −6.9  −1.2      deoA  Thymidine phosphorylase  2.1  −4.3  −6.9  −1.7      deoB  Deoxyribouratase, phosphopentomutase  −1.3  −4.2  −6.9  −3.6      nrdB  Ribonucleoside diphosphate reductase  −2.2  −2  −6.7  −3.2      fecB  Citrate-dependent iron transport  −2.2  −4.5  −6.7  −2.6      ycaR  Unknown  −1.7  −1.2  −6.5  −2.4      tnaL  Regulatory leader for tna operon  22.3  −4.2  −6.1  1.1      speD  S-Adenosylmethionine decarboxylase  −2.1  −2  −5.8  −3.5      rfbD  TDP-rhamnose synthetase  −3.8  1.1  −5.8  −1.2      ybaB  Unknown function  −2.2  −2.2  −5  −6.0          Recombinant proteina  rpoHb  Heat shockc        None  Folded  Misfolded      aRNA preparations were made from cell lines either not expressing recombinant protein or expressing folded or misfolded protein were compared topre-induction controls to determine fold changes in transcript levels.  bFold changes in transcript levels for indicated genes in cells containing a mutant rpoH gene when overexpressing misfolded protein.  cExpression data for the corresponding heat-shock genes as described by Richmond et al. (Richmond et al., 1999).  Heat-shock genes  ibpA  Chaperone  −1.5  −1.7  74.4  14.3  297.4    ibpB  Chaperone  1.4  2.7  40.0  10.4  327.2    yrfH  Ribosome-associated HSP  1.6  −3.2  28.3  2.8  51.3    yccV  Unknown  −1.9  −2.4  19.3  3.8  34.3    fxsA  Suppresses F exclusion of phage T7  −3.8  2.1  22.3  2.0  50.7    dnaK  Chaperone  1.8  3.2  16.6  3.8  58.5    htpG  Chaperone  −3.5  −2.6  13.2  3.0  33.8    clpP  Protease  2.6  −3.6  11.8  2.7  3.3    yhdN  Unknown  1.5  3.9  11.1  3.9  9.5    clpB  Protease  2.3  −1.3  9.7  7.0  36.5    hslV  Protease  1.1  1.7  7.4  3.5  16.2    mopA  Chaperone  1.9  −1.2  6.4  1.7  37.9    lon  Protease  −1.0  1.3  6.0  2.9  20.3    mopB  Chaperone  2.4  1.1  5.8  1.7  77.5    dnaJ  Chaperone  1.2  2.7  5.6  4.2  85.3    yrfG  Unknown  −1.0  −1.3  5.0  2.1  12.1    htpX  HSP, unknown  −2.5  −1.3  4.9  2.6  36.1    hslU  Protease  −1.0  −2.1  4.7  3.0  10.3    grpE  Chaperone  −1.4  1.5  3.9  2.6  24.1    yrfI  Chaperone  −1.5  −1.1  3.6  2.7  21.6    rrmJ  rRNA methylase  −1.2  1.5  3.2  3.0  9.1  Other induced  yagU  Unknown  2.6  2.4  17.4  3.1      yciS  Unknown  1.8  2.6  14.8  5.4      ybeD  Unknown  1.3  3.8  12.0  1.8      araE  Arabinose transport  1.2  3.1  11.6  17.7      yojH  Unknown  −1.9  5.2  9.7  −3.0      yejG  Unknown  1.6  3.0  7.3  5.0      exbB  Uptake of enterchelin  −1.6  −4.3  6.4  1.7      yhgI  Unknown  1.0  −1.0  5.3  2.2      proP  Proline transport  1.0  −1.0  5  5.1      kgtP  α-Ketoglutarate permease  −1.2  2.4  4.2  3.2    Downregulated  recR  Recombination and DNA repair  −3.3  −9.0  −17.9  −3.0      lamB  Maltose uptake  12.4  −4.8  −9.9  −5.7      glpD  Glycerol-3-phosphate dehydrogenase  −2.9  −2.3  −9  −8.1      yfiD  Unknown  5.3  −3.2  −8.6  −1.4      rbsC  d-Ribose transport  3.3  −5.9  −8.2  −3.0      glpF  Glycerol facilitator  −3.9  −3.2  −8  −10.3      yqjE  Unknown function  −1  −7.6  −7.7  1.1      ftsZ  Cell division; initiation of septation  −1.4  −8.0  −7.2  1.6      ycfN  Unknown function  −1.5  −1.3  −7.1  1.3      feoA  Ferrous iron uptake  2.1  −4.3  −7.1  1.1      ybjC  Unknown function  −1.5  −5.2  −6.9  −2.4      yccA  Unknown function  −3.6  −5.0  −6.9  −1.2      deoA  Thymidine phosphorylase  2.1  −4.3  −6.9  −1.7      deoB  Deoxyribouratase, phosphopentomutase  −1.3  −4.2  −6.9  −3.6      nrdB  Ribonucleoside diphosphate reductase  −2.2  −2  −6.7  −3.2      fecB  Citrate-dependent iron transport  −2.2  −4.5  −6.7  −2.6      ycaR  Unknown  −1.7  −1.2  −6.5  −2.4      tnaL  Regulatory leader for tna operon  22.3  −4.2  −6.1  1.1      speD  S-Adenosylmethionine decarboxylase  −2.1  −2  −5.8  −3.5      rfbD  TDP-rhamnose synthetase  −3.8  1.1  −5.8  −1.2      ybaB  Unknown function  −2.2  −2.2  −5  −6.0    View Large Table II. Fold change in cold-shock genes transcripts when misfolded protein is expressed   rpoH(+)  rpoH(−)  cspB  2.6  140  cspG  8.1  50  cspA  3.5  4.3  cspI  1.9  9.8    rpoH(+)  rpoH(−)  cspB  2.6  140  cspG  8.1  50  cspA  3.5  4.3  cspI  1.9  9.8  View Large Table III. Average solubility screen values   186 T.maritima proteinsa  Rep68    Soluble  Insoluble  Mixed  Screenb  Domainc  aAverage values given for cells expressing recombinant T.maritima proteins which are either soluble, insoluble or mixed solubility.  bAverage values given for 564 clones expressing truncated Rep68 proteins.  cValues for fragment of Rep68 showing improved solubility.  Relative β-galactosidase activity  96  680  280  280  3.1  Relative Ni-HRP absorbance  780  95  530  560  1700  Solubility score  300  2.9  83  3.7  550    186 T.maritima proteinsa  Rep68    Soluble  Insoluble  Mixed  Screenb  Domainc  aAverage values given for cells expressing recombinant T.maritima proteins which are either soluble, insoluble or mixed solubility.  bAverage values given for 564 clones expressing truncated Rep68 proteins.  cValues for fragment of Rep68 showing improved solubility.  Relative β-galactosidase activity  96  680  280  280  3.1  Relative Ni-HRP absorbance  780  95  530  560  1700  Solubility score  300  2.9  83  3.7  550  View Large Fig. 1. View largeDownload slide Expression of LCK and PLA. Proteins were expressed and fractionated as described previously. Equivalent fractional volumes were visualized by 4–20% SDS–PAGE. Lane 1, MW markers; lanes 2–5, PLA fractions; lanes 6–9, LCK fractions. Whole cell extracts after induction (lanes 3 and 7) show marked overexpression of the target proteins versus the preinduction controls (lanes 2 and 6). Overexpressed PLA is found almost exclusively in the soluble fraction versus the insoluble fraction (lane 4 versus lane 5), whereas LCK is found exclusively in the insoluble fraction (lane 8 versus lane 9). Fig. 1. View largeDownload slide Expression of LCK and PLA. Proteins were expressed and fractionated as described previously. Equivalent fractional volumes were visualized by 4–20% SDS–PAGE. Lane 1, MW markers; lanes 2–5, PLA fractions; lanes 6–9, LCK fractions. Whole cell extracts after induction (lanes 3 and 7) show marked overexpression of the target proteins versus the preinduction controls (lanes 2 and 6). Overexpressed PLA is found almost exclusively in the soluble fraction versus the insoluble fraction (lane 4 versus lane 5), whereas LCK is found exclusively in the insoluble fraction (lane 8 versus lane 9). Fig. 2. View largeDownload slide Summary of screening results for 18 T.maritima proteins with pre-determined expression characteristics. (A–C) Average relative β-galactosidase activity, Ni-HRP activity and the resulting solubility scores for 18 T.maritima proteins are shown. Expression characteristics for the 18 proteins were previously determined by SDS–PAGE of both soluble and insoluble fractions. (D) SDS–PAGE of representative T.maritima proteins showing soluble and insoluble fractions of cell lysates included in average values above. Lane 1, markers (sizes in kDa); lane 2, soluble lysate fraction (TM0414); lane 3, insoluble fraction (TM0414); lane 4, soluble lysate fraction (TM0688); lane 5, insoluble fraction (TM0688); lane 6, soluble lysate fraction (TM0564); lane 7, insoluble fraction (TM0564). Fig. 2. View largeDownload slide Summary of screening results for 18 T.maritima proteins with pre-determined expression characteristics. (A–C) Average relative β-galactosidase activity, Ni-HRP activity and the resulting solubility scores for 18 T.maritima proteins are shown. Expression characteristics for the 18 proteins were previously determined by SDS–PAGE of both soluble and insoluble fractions. (D) SDS–PAGE of representative T.maritima proteins showing soluble and insoluble fractions of cell lysates included in average values above. Lane 1, markers (sizes in kDa); lane 2, soluble lysate fraction (TM0414); lane 3, insoluble fraction (TM0414); lane 4, soluble lysate fraction (TM0688); lane 5, insoluble fraction (TM0688); lane 6, soluble lysate fraction (TM0564); lane 7, insoluble fraction (TM0564). Fig. 3. View largeDownload slide Relative β-galactosidase activity versus the relative Ni-HRP activity observed after expression of 186 T.maritima proteins in reporter strain. Classification of each protein as soluble, insoluble (A) or mixed (B) is based on SDS–PAGE performed on the soluble and insoluble lysates after the screen. Fig. 3. View largeDownload slide Relative β-galactosidase activity versus the relative Ni-HRP activity observed after expression of 186 T.maritima proteins in reporter strain. Classification of each protein as soluble, insoluble (A) or mixed (B) is based on SDS–PAGE performed on the soluble and insoluble lysates after the screen. Fig. 4. View largeDownload slide Alignment of secondary structure predictions and both predicted and identified domains of Rep68. Chou–Fasman secondary structure predictions of α-helical and β-sheet structures aligned with a Kyte–Doolittle plot of hydrophilicity based on the primary sequence of Rep68. Also, aligned below are blocks representing the relative size and position of the full-length Rep68 protein, the three predicted domains of Rep68 and the Rep68 domain identified by screening of randomly generated fragments of the rep68 gene. Solubility scores for the proteins are indicated. Fig. 4. View largeDownload slide Alignment of secondary structure predictions and both predicted and identified domains of Rep68. Chou–Fasman secondary structure predictions of α-helical and β-sheet structures aligned with a Kyte–Doolittle plot of hydrophilicity based on the primary sequence of Rep68. Also, aligned below are blocks representing the relative size and position of the full-length Rep68 protein, the three predicted domains of Rep68 and the Rep68 domain identified by screening of randomly generated fragments of the rep68 gene. Solubility scores for the proteins are indicated. 4 Present address: Syrrx Corporation, 10450 Science Center Drive, Suite 100, San Diego, CA 92121, USA 2 To whom correspondence should be addressed. E-mail: lesley@gnf.org The authors would like to thank Dan Giang, Tanya Shin, Juli Vincent and Dan McMullan for their help on expression profiling and protein purification. Cloning and expression of Thermotoga proteins was supported in part by grant GM62411-02. References Allen,S.P., Polazzi,J.O., Gierse,J.K. and Easton,A.M. ( 1992) J. Bacteriol. , 174, 6938–6947. Google Scholar Altschul,S.F., Madden,T.L., Schäffer,A.A., Zhang,J., Zhang,Z., Miller,W. and Lipman,D.J. ( 1997) Nucleic Acids Res. , 25, 3389–3402. Google Scholar Bae,W., Xia,B., Inuoye,M. and Severinov,K. ( 2000) Proc. Natl Acad. Sci. USA , 97, 7784–7789. Google Scholar Beckmann,R.P., Mizzen,L.A. and Welch,W.J. ( 1990) Science , 248, 850–854. Google Scholar Caldas,T., Binet,E., Bouloc,P., Costa,A., Desgres,J. and Richarme,G. ( 2000) J. Biol. Chem. , 275, 16414–16419. Google Scholar Caldas,T., Binet,E., Bouloc,P. and Richarme,G. ( 2000) Biochem. Biophys. Res. Commun. , 271, 714–718. Google Scholar Cashel,M., Gentry,D.R., Hernandez,V.J. and Vinella,D. (1996) In Neidhardt,F.C. (ed.), Escherichia coli and Salmonella Cellular and Molecular Biology. ASM Press, Washington DC, pp. 1458–1496. Google Scholar Gething,M.-J. ( 1997) Nature  (London), 388, 329–331. Google Scholar Goloubinoff,P., Mogk,A., Zui,A.P.B., Tomoyasu,T. and Bukau,B. ( 1999) Proc. Natl Acad. Sci. USA , 96, 12732–12737. Google Scholar Hartl,F.U. ( 1996) Nature  (London), 381, 571–580. Google Scholar Jiang,W., Hou,Y. and Inouye,M. ( 1997) J. Biol. Chem. , 272, 196–202. Google Scholar Korber,P., Stahl,J.M., Nierhaus,K.H. and Bardwell,J.C.A. ( 2000) EMBO J. , 19, 741–748. Google Scholar Liberek,K. and Georgopoulos,C. ( 1993) Proc. Natl Acad. Sci. USA , 90, 11019–11023. Google Scholar Lockhart,D.J., Dong,H., Byrne,M.C., Follettie,M.T., Gallo,M.V., Chee,M.S., Mittmann,M., Wang,C., Kobayashi,M., Horton,H. and Brown,E.L. ( 1996) Nat. Biotechnol. , 14, 1675–1680. Google Scholar Maxwell,K.L., Mittermaier,A.K., Forman-Kay,J.D., and Davidson,A.R. ( 1999) Protein Sci. , 8, 1908–1911. Google Scholar McCarty,J.S., Rudiger,S., Schonfeld,H.-J., Schneider-Mergener,J., Nakahigashi,K., Yura,T. and Bukau,B. ( 1996) J. Mol. Biol. , 256, 829–837. Google Scholar Mogk,A., Tomoyasu,T., Goloubinoff,P., Rudiger,S., Roder,D. Langen,H. and Bukau,B. ( 1999) EMBO J. , 18, 6934–6949. Google Scholar Parsell,D.A. and Sauer,R.T. ( 1989) Genes Dev. , 3, 1226–1232. Google Scholar Puglisi,J.D., Blanchard,S.C. and Green,R. ( 2000) Nat. Struct. Biol. , 7, 855–861. Google Scholar Richmond,C.S., Glasner,J.D., Mau,R., Hongfan,J. and Blattner,F.R. ( 1999) Nucleic Acids Res. , 27, 3821–3835. Google Scholar Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) In Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, p. 1735. Google Scholar Thomas,J.G. and Baneyx,F. ( 1998) J. Bacteriol. , 180, 5165–5172. Google Scholar Veinger,L., Diamant,S., Buchner,J. and Goloubinoff,P. ( 1998) J. Biol. Chem. , 273, 11032–11037. Google Scholar Waldo,G.S., Standish,B.M., Berendzen,J. and Terwilliger,T.C. ( 1999) Nat. Biotechnol. , 17, 691–695. Google Scholar Wang,N., Yamanake,K. and Inouye,M. ( 1999) J. Bacteriol. , 181, 1603–1609. Google Scholar Wigley,W.C., Stidham,R.D., Smith,N.M., Hunt,J.F. and Thomas,P.J. ( 2001) Nat. Biotechnol. , 19, 131–135. Google Scholar Wodicka,L., Dong,H., Mittmann,M., Ho,M.H., and Lockhart,D.J. (0 1997) Nat. Biotechnol. , 15, 1359–1367. Google Scholar © Oxford University Press http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Protein Engineering, Design and Selection Oxford University Press

Gene expression response to misfolded protein as a screen for soluble recombinant protein

Loading next page...
 
/lp/oxford-university-press/gene-expression-response-to-misfolded-protein-as-a-screen-for-soluble-24uaaDAuNs

References (38)

Publisher
Oxford University Press
Copyright
© Oxford University Press
ISSN
1741-0126
eISSN
1741-0134
DOI
10.1093/protein/15.2.153
Publisher site
See Article on Publisher Site

Abstract

Abstract Proper protein folding is key to producing recombinant proteins for structure determination. We have examined the effect of misfolded recombinant protein on gene expression in Escherichia coli. Comparison of expression patterns indicates a unique set of genes responding to translational misfolding. The response is in part analogous to heat shock and suggests a translational component to the regulation. We have further utilized the expression information to generate reporters responsive to protein misfolding. These reporters were used to identify properly folded recombinant proteins and to create soluble domains of insoluble proteins for structural studies. Introduction Structural genomics efforts are an important element of evaluating gene function. Protein expression and purification are key processes in these studies, and are often limited by the ability to produce properly folded recombinant protein. Escherichia coli is a common expression host that often makes misfolded protein when obliged to overproduce non-native gene products. It is easy to overlook that this workhorse is itself an evolved biological organism reacting to a stress, namely overproduction of a foreign gene product. Gene expression in response to protein misfolding is key to understanding the reaction to this stress and potential solutions to the practical aspects of recombinant protein expression. A variety of environmental stresses cause protein denaturation resulting in a classical heat-shock response. Unfolded protein is the stimulus for the change in gene expression (Parsell and Sauer, 1989). It is estimated that 10–15% of the proteins in E.coli are at least partially denatured under heat stress (Mogk et al., 1999). DnaK interacts with these proteins and initiates a refolding pathway involving DnaJ and GrpE (Beckmann et al., 1990; Hartl, 1996; Gething, 1997; Goloubinoff et al., 1999). DnaK is also a central transcription-regulator by negatively influencing the transcription factor RpoH (Liberek and Georgopoulos, 1993; McCarty et al., 1996). DnaK–RpoH interaction provides a feedback mechanism and is the proposed switch by which the cell senses the level of unfolded protein and can respond through appropriate gene expression. A major source of unfolded proteins in vivo is the nascent chain of translating genes. Gene expression in response to translational stress is related to the heat-shock response (Parsell and Sauer, 1989), but has not been closely examined. Genome arrays used to study global gene expression in E.coli at elevated temperature (Richmond et al., 1999) show changes in a large number of transcripts. To study translational misfolding rather than heat stress, we used recombinant expression of folded and misfolded proteins as a stimulus. Our results indicate a unique set of genes respond to translational misfolding. The nature of these genes implies that many known heat-shock and chaperone proteins, as well as ribosome-associated proteins suggestive of translational regulation, are induced to deal with translational stress. We have applied these results in a practical way by developing a screen to improve our ability to generate recombinant proteins for structural studies. Misfolded protein often appears as insoluble aggregates when overexpressed in E.coli. Several recent approaches have been used to detect soluble recombinant protein. They utilize protein fusion to a reporter gene such as CAT (Maxwell et al., 1999), GFP (Waldo et al., 1999), or lacZα (Wigley et al., 2001) as a measure of the amount of soluble protein expressed. Though convenient, these reporters have the potential to be biased by the nature of the fusion. The reporter portion of the fusion may alter the solubility of the target protein, either positively or negatively, giving unexpected results when expressing in the absence of the reporter fusion. By utilizing a sensory and regulatory system already existing in the host, we have created a general screen for detecting misfolded, and typically insoluble, recombinant proteins without the requirement of a direct protein fusion. We have also demonstrated the utility of this approach combined with mutagenesis to create soluble fragments of recombinant proteins in E.coli. The approach provides a general means of providing folded domains for structural studies. Materials and methods Cloning Clones expressing properly folded or misfolded human proteins were obtained from the GeneStorm collection (Invitrogen). Clones containing the Unigene accession numbers L35545, U18291, M94856, M22146, D87116, M63167, M68520, M60527, M36881, M36981, U35003, S79522, X73460, D14520, U14968, M86400 were provided in the pBADThio vector to provide arabinose-inducible expression. Thermotoga maritima genes were amplified from genomic DNA and cloned into the expression vector pMH1 which encodes a 12 amino acid N-terminal tag containing a 6X-histidine repeat for purification and detection. Reporter vectors were constructed by insertion of a PCR amplifer of 300 bp upstream of the ibpAB, ybeD, yhgI or yrfGHI genes upstream of β-galactosidase in a pACYC184 derivative. Rep68 was cloned from ATCC 68066 containing the entire genome of the human adeno-associated virus 2 (AAV2). Putative domains comprised of bases 1–646, 647–1456 and 1457–1611 were amplified from the full-length template and cloned into pMH1. The above template was also used in amplifications of the full-length gene for fragmentation. Two micrograms of the Rep68 amplifer were used in each of five fragmentation reactions containing 1, 0.1, 0.01, 0.001 or 0 U DNase I (Boeringer Mannheim) as well as Pfu polymerase and dNTPs. Reactions were set up on ice with the DNase added immediately prior to temperature cycling in an MJ Research thermocycler according to the following: 10 min at 25°C, 15 min at 95°C and 30 min at 72°C. Each reaction was run on a 1% agarose gel and fragments corresponding to 1600–1000, 1000–850, 850–600 and 600–300 bp were extracted. Each pool was used as above for blunt cloning and ligation into pMH1 as above and introduced into the reporter cell line HK 57 for screening. Thermotoga proteins used for expression studies evaluating proteins of known expression characteristics were cloned into pMH1 as described above. Coding regions were introduced for: TM0560, TM0414, TM0574, TM0703, TM0554, TM0556 (soluble expression); TM0688, TM0633, TM0712, TM0343, TM0218, TM0294 (mixed expression); TM0289, TM0564, TM0540, TM0425, TM0731, TM0413 (insoluble expression). Cell growth and protein expression Escherichia coli strains MG1655 (F−lam rph1) and KY1429 (F-araD139 Δ(argF-lac)169 lam flhD5301 fruA25 relA1 rpsL150 zhh50::Tn10 rpoH606(ts) deoC1) were transformed with expression plasmids encoding M36881 [human lymphocyte-specific protein tyrosine kinase (LCK)] or M86400 [human phospholipase A2 (PLA)] for expression profiling. Cells were cultured at 37°C in Luria broth (LB) containing ampicillin. Protein expression was induced by the addition of l-arabinose to a final concentration of 0.1% for 1 h. KY1429 cells were cultured as above except initial growth was performed at 32°C followed by a shift to 42°C for non-permissive expression of rpoH606. Top10 cells (F−mcrA Δ(mrr-hsdRMS-mcrBC) ϕ80lacZΔM15 ΔlacX74 deoR recA1 araD139 Δ(ara-leu)7697 galU galK rpsL endA1 nupG) containing the ibpAB promoter fusion (pHK57), were transformed with expression constructs listed above. β-Galactosidase assays were performed essentially as described by Miller (Sambrook et al., 1989). Fractionation of soluble and insoluble proteins was performed by centrifugation. Cultures were resuspended in 50 mM Tris pH 7.9, 50 mM NaCl, 1 mM MgCl2, 3 mM methionine and sonicated for 2 min on ice. Cell debris and insoluble protein aggregates were pelleted by centrifugation at 3000 g for 15 min. The soluble fraction was removed and the pellets resuspended in an equivalent volume of lysis buffer. Probe preparation and hybridization and analysis of labeled mRNA Labeled mRNA was prepared and hybridized to an E.coli whole genome array (Affymetrix) essentially as described previously (Lockhart et al., 1996; Wodicka et al., 1997). This gene chip contains 25-mer oligonucleotide probes for each of the 4290 known E.coli genes. Standard Affymetrix GeneChip analysis software was used to measure individual gene expression and to perform pairwise comparison of gene expression levels for pre-induction and post-induction samples. Comparisons of changes in gene expression for properly folded and misfolded genes were analyzed for individual gene probe sets. Microplate solubility screening Ninety-six-well microplates containing 200 μl of LB with 100 μg/ml ampicillin and 34 μg/ml chloramphenicol were inoculated with single colonies from above and grown overnight with shaking at 37°C. Overnight cultures were used to inoculate 200 μl of the same media and incubated at 37°C until reaching an average OD600 of 0.5. Cultures were induced with a final concentration of 0.2% arabinose. After 30 min, a cocktail of ceftriaxone and cefotaxime was added to each well to a final concentration of 10 μg/ml of each and the plates were incubated for an additional 1.5 h. These antibiotics provide a convenient method of lysis in microplate format without the need to add detergents that might solublize misfolded protein. Cultures were harvested after 2 h total of induction by centrifugation at maximum speed for 15 min to pellet cell debris on the bottom of the wells. The soluble lysate was then separated by transferring 25 μl into one set of clean microplates for β-galactosidase activity screens and 75 μl into Nunc Maxisorp ELISA plates for Ni-HRP screening. β-Galactosidase activity screening of lysates was performed using a variation of the Miller protocol (Maxwell et al., 1999). A 50 μl aliquot of 4× Z-buffer and 50 μl of 4× ONPG were added to microplates containing 25 μl of soluble lysate. After development of yellow color in positive control wells, the reaction was quenched with 75 μl of 1 M Na2CO3 pH 8. The A420, A550 and reaction times were recorded and used along with the OD600 data to calculate β-galactosidase activity (Maxwell et al., 1999). Ni-HRP screening was performed similar to an ELISA. A 75 μl aliquot of lysate plus 25 μl TBS was bound overnight at 4°C to a microtiter plate and blocked with 1% (w/v) BSA in TBS for 4 h at 25°C. Plates were then washed 3× with TBST and 100 μl of Ni-HRP conjugate (KPL Labs) was added at a dilution of 1:2500 and incubated 1 h at 25°C. The plates were then washed with TBST and 100 μl of the HRP substrate (KPL Labs) was added and color was allowed to develop until the positive control well was deep blue. The reaction was quenched with 100 μl of 1 N HCl and the A420 determined. Solubility scores were calculated by weighting the Ni-HRP A420 readings such that the experimental mean was one order of magnitude greater than the mean of the β-galactosidase activity scores, then dividing the Ni-HRP absorbance by the β-galactosidase activity. This calculation was found empirically to provide good distinction between soluble and insoluble proteins. Results Analysis of gene expression To examine gene expression as a result of misfolded protein, representative genes were cloned as fusion proteins to thioredoxin under control of the tightly regulated arabinose promoter. PLA is almost entirely soluble, as determined by cell lysis and fractionation by centrifugation (Figure 1). Further evidence of proper folding of this protein was obtained through dynamic light scattering of purified protein and the ability to crystallize it from a single affinity purification step (unpublished data). Under equivalent expression conditions, LCK is expressed almost exclusively as insoluble protein. Both proteins were expressed at sufficient levels to be the predominant translation product. mRNA preparations from induced and non-induced cultures were prepared and used to probe for gene expression. Recombinant protein expression within E.coli is predicted to cause a substantial change in gene expression. Indeed, a comparison of gene expression with a pre-induction control shows 6% of total genes with >3-fold differences in expression regardless of whether the expressed protein is soluble or insoluble. In the case of insoluble recombinant protein, 27 genes show >10-fold changes in expression, as compared to 10 genes in the case of the soluble recombinant protein. A comparison of the two profiles identifies 52 genes listed in Table I showing >3-fold changes, that are unique to the insoluble case. These genes, then, are likely responsive to misfolded protein in the cell and may play a role within E.coli in dealing with this translational stress. The heat-shock transcription factor RpoH is normally repressed by interaction with the chaperone protein DnaK. In the presence of misfolded protein, DnaK binds to that protein thereby allowing RpoH to stimulate transcription of heat-shock promoters (Liberek and Georgopoulos, 1993). Upstream regions of many of the induced genes in Table I show the presence of RpoH-dependant promoter sequences. Further evidence of the important role played by RpoH is provided by expression profiling results performed from an rpoH606 mutant (KY1429) expressing misfolded LCK protein compared to a non-expressing control. A strikingly different expression profile is seen in the case of the rpoH606 mutant (Tables I and II). The majority of the genes induced by the misfolded protein in the wild-type strain are poorly induced in the rpoH606 mutant indicating that they are directly or indirectly under the control of this transcription factor. Induction of heat-shock genes Not surprisingly, many of the genes induced by translational misfolding have known chaperone activity. These include the well characterized dnaJ, dnaK and grpE genes. The corresponding proteins interact as a complex with misfolded or denatured protein in an ATP-dependant repair process. Likewise, mopAB genes forming the GroELS folding repair complex are induced under translational misfolding conditions. IbpAB are small heat-shock polypeptides associated with inclusion body aggregates of recombinant protein (Allen et al., 1992). Whereas they do not appear to behave as folding chaperones directly, they bind misfolded protein and interact with the DnaJK GrpE proteins as a chaperone system (Thomas and Baneyx, 1998). Hsp33, the gene product of the yrfI gene was recently identified as a chaperone protein responsive to oxidizing conditions (Veinger et al., 1998). Genes implicated in degradation of denatured protein are also induced by translational misfolding. The lon, clpBP and hslUV protease genes are expressed at increased levels. Under normal cell growth these proteases serve an important recycling function. Insoluble aggregates are relatively resistant to proteolysis and this recycling pathway is ineffective for recombinant protein expression. Induction of ribosome-associated genes Other heat-shock genes associated with the ribosome are induced under conditions of translational misfolding. Hsp15 (yrfH) binds RNA (Sambrook et al., 1989) and is associated with free 50S ribosomal subunits containing a nascent polypeptide chain (Korber et al., 2000). Heat shock also increases the level of Hsp15-binding implying increased dissociation of 50S and 30S subunits. Further suggestion of ribosomal dissociation comes from the induction of ftsJ (rrmJ). The ftsJ gene product is an RNA methylase specific for 23S rRNA only when contained in the 50S ribosomal subunit (Caldas et al., 2000a,b; Puglisi et al., 2000). This enzyme methylates 23S rRNA at position 2552 located within the peptidyl transferase center of the ribosome (Caldas et al., 2000a). Mutants in ftsJ lack methylation of 23S rRNA and show up to 65% decrease in ribosomal activity corresponding to dissociation of the 50S and 30S subunits (Caldas et al., 2000b). Particularly striking in the rpoH mutant is the large increase in transcripts of the cold-shock proteins (CSPs). Table II shows the response of CSP transcripts to misfolded protein in the rpoH mutant and the wild-type rpoH strain. These genes are not affected by heat shock (9), but are associated with a transient halt of translation. CSPs are RNA-binding proteins which act as chaperones for untranslated message (Jiang et al., 1997; Wang et al., 1999) and provide anti-termination activity (Bae et al., 2000). Increased expression of CSPs under conditions that reduce chaperone expression (rpoH606) is an indication of paused translation. Taken together, these results suggest a translational regulatory response to misfolded protein. Such regulation might involve rRNA demethylation, as a consequence of translational misfolding. This hypothesis is an interesting regulatory mechanism currently under investigation. Other induced genes yccV, yhdN and yrfG have been shown to increase expression under heat-shock conditions but are of unknown function. In addition to these known heat-shock genes, yagU, yciS, ybeD, yejG and yhgI show increased expression. Most of these proteins are relatively small and generally acidic. One speculation is that some of these proteins perform a similar role to IbpAB in the direct recognition and sequestering of misfolded protein. However, only IbpAB have been associated with misfolded and aggregated protein. Induction levels of ibpAB are much higher and these other proteins may be present at lower levels. Interestingly, knockout mutations of ibpAB have relatively little affect on cell growth and viability (Thomas and Baneyx, 1998; data not shown) suggesting some functional redundancy within the cell. Genetic reporter of protein folding To confirm the profiling results and facilitate experimentation with a larger number of recombinant proteins, we cloned the promoter regions from ibpAB, ybeD, yhgI and yrfGHI into a β-galactosidase reporter vector. In each case, increased β-galactosidase activity was observed when expression of the misfolded protein LCK was induced whereas the folded protein PLA showed no increase in activity. These results were further extended using a set of eight misfolded proteins and six properly folded proteins co-expressed in the presence of the ibpAB-promoter β-galactosidase fusion. In each case, increased β-galactosidase activity corresponded to expression of misfolded protein (data not shown). A more detailed characterization is shown below. The response observed, then, appears to be a general result of protein misfolding rather than a specific response to any particular protein. These reporters provide a simple means of identifying misfolded protein through a sensitive enzymatic assay and the ibpAB promoter fusion was chosen as the reporter for further studies. ELISA-like assay for soluble protein For identifying protein derivatives that have improved folding properties in a recombinant environment, we also developed an ELISA-like assay compatible with high-throughput screening instrumentation. To evaluate soluble protein levels in a high-throughput system, non-denatured cell lysates must be prepared using conditions compatible with rapid screening in microplates. In lieu of the detergent or organic lysis, we added an antibiotic cocktail to each well to induce lysis. Soluble protein fractions were removed, bound to microtiter plates, and recombinant protein detected via binding of a Ni-HRP conjugate to a 6X-histidine N-terminal fusion. It should be noted that the His-tag may not be uniformly accessible among recombinant proteins. A negative Ni-HRP response, therefore, may not be indicative of an absence of soluble protein, but the protein fold may occlude access to the His-tag. However, we have not observed this to be a common problem. This assay, then, provides a measure of the levels of soluble recombinant protein without the need to run an SDS gel and in a form that is compatible with a HT-screen and the β-galactosidase assay. Testing proteins with pre-determined expression characteristics As part of our effort aimed at cloning, expressing and characterizing the total proteome of T.maritima, we tested the efficacy of the reporter on a set of 18 T.maritima proteins (six soluble, six insoluble and six mixed solubility). To optimize assay parameters, strains were arrayed in 96-well plates and assayed in triplicate at three induction levels (0.02, 0.2 and 2% arabinose) and at four post-induction time points for addition of the lysis-promoting antibiotics (t=0, 30, 60 and 120 min after addition of arabinose.) Figure 2 shows the averaged results for triplicate plates (soluble, insoluble and mixed) for the 0.2% arabinose induction. Both the insoluble and the mixed pools showed >4-fold higher β-galactosidase activity than the soluble pool. Conversely, the soluble pool showed a >10-fold higher response in the Ni-HRP assay opposed to the insoluble pool. The mixed pool, comprised of proteins expressed approximately equally in both soluble and insoluble fractions, showed Ni-HRP binding approximately half the intensity of the soluble pool. Although either lack of β-galactosidase or presence of Ni-HRP activity alone could be used as a measure of soluble protein, we chose a ratio of the two activities, a more effective and convenient screen. Testing proteins with unknown expression characteristics We next applied this screen to a large set of proteins with unknown folding properties. We performed the screen under the optimal conditions noted above on 186 T.maritima proteins not previously characterized for expression. The results of this screen are summarized in Table III. SDS–PAGE of eluates from nickel-chelating resin and the dissolved insoluble fractions for each clone was performed along with corresponding β-galactosidase activity, Ni-HRP response and solubility scores for 186 clones. Based on the results of the gels, 57 clones did not overexpress a visible protein band, 62 clones expressed predominantly soluble protein, 27 expressed predominantly insoluble aggregates and 46 expressed approximately equally in both soluble and insoluble fractions. A comparison of β-galactosidase activity to the Ni-HRP assay is shown in Figure 3. Points are categorized by SDS gel analysis of the soluble and insoluble protein fractions. The screen positively identified 54 of 62 (87%) soluble proteins. Seven of the eight remaining proteins that were soluble according to the gels had low Ni-HRP assays, most likely due to inaccessibility of the His-tag in these fusion proteins. Taken alone, the β-galactosidase activity measurement identified 22 of 27 (81%) insoluble proteins. Those proteins showing partial solubility showed variable solubility scores, suggesting partial folding is inducing β-galactosidase through the reporter. This assay, then, provides an effective and convenient means of classifying folding characteristics. Identification of soluble protein domains The true utility of this system lies in the ability to identify variants of full-length gene products, either mutants or domains, based on improved properties. For structural and biochemical studies, we tested the ability of this screen to identify soluble fragments of Rep68 (GI: 209617), an adeno-associated virus non-structural protein possessing various activities related to the integration of the viral genome into target DNA. This protein previously had been found to express predominantly as unfolded aggregates in E.coli. We performed both a random approach and a rational approach based on selection of domains with regard to homology. Three domains of Rep68 were selected after an RPS-BLAST search (Altschul et al., 1997) identified an internal domain possessing homology to a parvovirus non-structural protein, NP-1. This information, combined with a Kyte–Doolittle hydropathy plot (Figure 4), was used to assign the 5′ and 3′ cut-offs for each domain. The remaining N-terminal and C-terminal residues comprised the other two domains and did not possess significant homology to any other proteins in the database. Random fragments of Rep68 were also generated for screening by DNase fragmentation. The three predicted domains and randomly generated fragments were screened to identify soluble fragments of Rep68. None of the three predicted domains were identified as soluble (Figure 4). Concurrently, 564 randomly generated fragments of rep68 were also screened. One fragment returned a significantly high solubility score (Table III). This clone was verified by large-scale expression and showed expression in both the soluble and insoluble fractions. Subsequent sequencing of the identified clone verified that it was comprised of a fragment of rep68 corresponding to amino acids 1–95 (Figure 4). The identified fragment showed substantial improvement in solubility over the full-length protein and is being tested in crystallization trials. Discussion Most gene expression studies of in vivo protein folding have focused on denaturation as a result of environmental stress. This response is essential in vivo to deal with ever-changing environmental and non-ideal growth conditions. Translational folding issues are equally important to the cell since every protein as it is being translated is essentially in an unfolded state. Expression of unnatural proteins, either through recombinant means or mutation is a `stress' in itself. We show that the cellular response to translational misfolding, like heat shock, involves many known chaperone genes with a clear inference how these gene products are involved in the folding of the nascent polypeptide. In addition, other non-heat-shock genes and genes of unknown function are induced. These too may be involved in the folding process. Our results suggest both transcriptional and translational regulation. The DnaK–RpoH interaction is well characterized and appears to be the major regulator of the transcriptional response. The altered expression of genes implicated in translational stalling and ribosomal dissociation is intriguing and implies that these effects might be a result of translationally misfolded protein. These genes include yrfH which associates with the 50S ribosomal subunit (Korber et al., 2000). cspABGI are induced in rpoH606 suggesting translational stalling in the absence of induced chaperones. Also included is ftsJ, which is known to methylate the 23S rRNA of 50S subunits resulting in higher affinity of the two subunits for each other (Caldas et al., 2000a,b). Ribosome structure shows that the location of the methylation site of ftsJ, position 2552 of the 23S rRNA, is intriguingly close to the peptidyl transferase center (Puglisi et al., 2000) making it an obvious potential regulator mechanism for a ribosomal sensor of misfolded protein. Such a ribosomal sensor is not unprecedented as demonstrated by the well characterized stringent response to uncharged tRNAs during translation (Cashel et al., 1996). Ribosomal stalling provides a mechanism to allow time for chaperone synthesis and recruitment thereby preventing irreversible aggregation. In this way, the cell would retain an additional salvage pathway where the emerging protein was held in the relatively protected environment of the translating ribosome until sufficient chaperones could be recruited. The differentially regulated genes identified provide a valuable opportunity to create novel reporters of the folding state of cellular proteins as a whole and overexpressed, recombinant proteins in particular. Our reporter assay differs from others recently described by not relying on direct coupling of the reporter gene to the target, thereby limiting potential interference by the reporter. The combination of the Ni-HRP and β-galactosidase assays provides an effective means of assaying soluble recombinant proteins in a high-throughput way. We have extended this system to identify mutants and truncations of single gene products as a strategy to identify soluble domains of otherwise misfolded, aggregated proteins. Using this approach, we have identified soluble fragments of Rep68 and anticipate that this assay will provide a general means of isolating recombinant protein suitable for structure/function work. Table I. Fold change in gene expression for genes unique to misfolded response       Recombinant proteina  rpoHb  Heat shockc        None  Folded  Misfolded      aRNA preparations were made from cell lines either not expressing recombinant protein or expressing folded or misfolded protein were compared topre-induction controls to determine fold changes in transcript levels.  bFold changes in transcript levels for indicated genes in cells containing a mutant rpoH gene when overexpressing misfolded protein.  cExpression data for the corresponding heat-shock genes as described by Richmond et al. (Richmond et al., 1999).  Heat-shock genes  ibpA  Chaperone  −1.5  −1.7  74.4  14.3  297.4    ibpB  Chaperone  1.4  2.7  40.0  10.4  327.2    yrfH  Ribosome-associated HSP  1.6  −3.2  28.3  2.8  51.3    yccV  Unknown  −1.9  −2.4  19.3  3.8  34.3    fxsA  Suppresses F exclusion of phage T7  −3.8  2.1  22.3  2.0  50.7    dnaK  Chaperone  1.8  3.2  16.6  3.8  58.5    htpG  Chaperone  −3.5  −2.6  13.2  3.0  33.8    clpP  Protease  2.6  −3.6  11.8  2.7  3.3    yhdN  Unknown  1.5  3.9  11.1  3.9  9.5    clpB  Protease  2.3  −1.3  9.7  7.0  36.5    hslV  Protease  1.1  1.7  7.4  3.5  16.2    mopA  Chaperone  1.9  −1.2  6.4  1.7  37.9    lon  Protease  −1.0  1.3  6.0  2.9  20.3    mopB  Chaperone  2.4  1.1  5.8  1.7  77.5    dnaJ  Chaperone  1.2  2.7  5.6  4.2  85.3    yrfG  Unknown  −1.0  −1.3  5.0  2.1  12.1    htpX  HSP, unknown  −2.5  −1.3  4.9  2.6  36.1    hslU  Protease  −1.0  −2.1  4.7  3.0  10.3    grpE  Chaperone  −1.4  1.5  3.9  2.6  24.1    yrfI  Chaperone  −1.5  −1.1  3.6  2.7  21.6    rrmJ  rRNA methylase  −1.2  1.5  3.2  3.0  9.1  Other induced  yagU  Unknown  2.6  2.4  17.4  3.1      yciS  Unknown  1.8  2.6  14.8  5.4      ybeD  Unknown  1.3  3.8  12.0  1.8      araE  Arabinose transport  1.2  3.1  11.6  17.7      yojH  Unknown  −1.9  5.2  9.7  −3.0      yejG  Unknown  1.6  3.0  7.3  5.0      exbB  Uptake of enterchelin  −1.6  −4.3  6.4  1.7      yhgI  Unknown  1.0  −1.0  5.3  2.2      proP  Proline transport  1.0  −1.0  5  5.1      kgtP  α-Ketoglutarate permease  −1.2  2.4  4.2  3.2    Downregulated  recR  Recombination and DNA repair  −3.3  −9.0  −17.9  −3.0      lamB  Maltose uptake  12.4  −4.8  −9.9  −5.7      glpD  Glycerol-3-phosphate dehydrogenase  −2.9  −2.3  −9  −8.1      yfiD  Unknown  5.3  −3.2  −8.6  −1.4      rbsC  d-Ribose transport  3.3  −5.9  −8.2  −3.0      glpF  Glycerol facilitator  −3.9  −3.2  −8  −10.3      yqjE  Unknown function  −1  −7.6  −7.7  1.1      ftsZ  Cell division; initiation of septation  −1.4  −8.0  −7.2  1.6      ycfN  Unknown function  −1.5  −1.3  −7.1  1.3      feoA  Ferrous iron uptake  2.1  −4.3  −7.1  1.1      ybjC  Unknown function  −1.5  −5.2  −6.9  −2.4      yccA  Unknown function  −3.6  −5.0  −6.9  −1.2      deoA  Thymidine phosphorylase  2.1  −4.3  −6.9  −1.7      deoB  Deoxyribouratase, phosphopentomutase  −1.3  −4.2  −6.9  −3.6      nrdB  Ribonucleoside diphosphate reductase  −2.2  −2  −6.7  −3.2      fecB  Citrate-dependent iron transport  −2.2  −4.5  −6.7  −2.6      ycaR  Unknown  −1.7  −1.2  −6.5  −2.4      tnaL  Regulatory leader for tna operon  22.3  −4.2  −6.1  1.1      speD  S-Adenosylmethionine decarboxylase  −2.1  −2  −5.8  −3.5      rfbD  TDP-rhamnose synthetase  −3.8  1.1  −5.8  −1.2      ybaB  Unknown function  −2.2  −2.2  −5  −6.0          Recombinant proteina  rpoHb  Heat shockc        None  Folded  Misfolded      aRNA preparations were made from cell lines either not expressing recombinant protein or expressing folded or misfolded protein were compared topre-induction controls to determine fold changes in transcript levels.  bFold changes in transcript levels for indicated genes in cells containing a mutant rpoH gene when overexpressing misfolded protein.  cExpression data for the corresponding heat-shock genes as described by Richmond et al. (Richmond et al., 1999).  Heat-shock genes  ibpA  Chaperone  −1.5  −1.7  74.4  14.3  297.4    ibpB  Chaperone  1.4  2.7  40.0  10.4  327.2    yrfH  Ribosome-associated HSP  1.6  −3.2  28.3  2.8  51.3    yccV  Unknown  −1.9  −2.4  19.3  3.8  34.3    fxsA  Suppresses F exclusion of phage T7  −3.8  2.1  22.3  2.0  50.7    dnaK  Chaperone  1.8  3.2  16.6  3.8  58.5    htpG  Chaperone  −3.5  −2.6  13.2  3.0  33.8    clpP  Protease  2.6  −3.6  11.8  2.7  3.3    yhdN  Unknown  1.5  3.9  11.1  3.9  9.5    clpB  Protease  2.3  −1.3  9.7  7.0  36.5    hslV  Protease  1.1  1.7  7.4  3.5  16.2    mopA  Chaperone  1.9  −1.2  6.4  1.7  37.9    lon  Protease  −1.0  1.3  6.0  2.9  20.3    mopB  Chaperone  2.4  1.1  5.8  1.7  77.5    dnaJ  Chaperone  1.2  2.7  5.6  4.2  85.3    yrfG  Unknown  −1.0  −1.3  5.0  2.1  12.1    htpX  HSP, unknown  −2.5  −1.3  4.9  2.6  36.1    hslU  Protease  −1.0  −2.1  4.7  3.0  10.3    grpE  Chaperone  −1.4  1.5  3.9  2.6  24.1    yrfI  Chaperone  −1.5  −1.1  3.6  2.7  21.6    rrmJ  rRNA methylase  −1.2  1.5  3.2  3.0  9.1  Other induced  yagU  Unknown  2.6  2.4  17.4  3.1      yciS  Unknown  1.8  2.6  14.8  5.4      ybeD  Unknown  1.3  3.8  12.0  1.8      araE  Arabinose transport  1.2  3.1  11.6  17.7      yojH  Unknown  −1.9  5.2  9.7  −3.0      yejG  Unknown  1.6  3.0  7.3  5.0      exbB  Uptake of enterchelin  −1.6  −4.3  6.4  1.7      yhgI  Unknown  1.0  −1.0  5.3  2.2      proP  Proline transport  1.0  −1.0  5  5.1      kgtP  α-Ketoglutarate permease  −1.2  2.4  4.2  3.2    Downregulated  recR  Recombination and DNA repair  −3.3  −9.0  −17.9  −3.0      lamB  Maltose uptake  12.4  −4.8  −9.9  −5.7      glpD  Glycerol-3-phosphate dehydrogenase  −2.9  −2.3  −9  −8.1      yfiD  Unknown  5.3  −3.2  −8.6  −1.4      rbsC  d-Ribose transport  3.3  −5.9  −8.2  −3.0      glpF  Glycerol facilitator  −3.9  −3.2  −8  −10.3      yqjE  Unknown function  −1  −7.6  −7.7  1.1      ftsZ  Cell division; initiation of septation  −1.4  −8.0  −7.2  1.6      ycfN  Unknown function  −1.5  −1.3  −7.1  1.3      feoA  Ferrous iron uptake  2.1  −4.3  −7.1  1.1      ybjC  Unknown function  −1.5  −5.2  −6.9  −2.4      yccA  Unknown function  −3.6  −5.0  −6.9  −1.2      deoA  Thymidine phosphorylase  2.1  −4.3  −6.9  −1.7      deoB  Deoxyribouratase, phosphopentomutase  −1.3  −4.2  −6.9  −3.6      nrdB  Ribonucleoside diphosphate reductase  −2.2  −2  −6.7  −3.2      fecB  Citrate-dependent iron transport  −2.2  −4.5  −6.7  −2.6      ycaR  Unknown  −1.7  −1.2  −6.5  −2.4      tnaL  Regulatory leader for tna operon  22.3  −4.2  −6.1  1.1      speD  S-Adenosylmethionine decarboxylase  −2.1  −2  −5.8  −3.5      rfbD  TDP-rhamnose synthetase  −3.8  1.1  −5.8  −1.2      ybaB  Unknown function  −2.2  −2.2  −5  −6.0    View Large Table II. Fold change in cold-shock genes transcripts when misfolded protein is expressed   rpoH(+)  rpoH(−)  cspB  2.6  140  cspG  8.1  50  cspA  3.5  4.3  cspI  1.9  9.8    rpoH(+)  rpoH(−)  cspB  2.6  140  cspG  8.1  50  cspA  3.5  4.3  cspI  1.9  9.8  View Large Table III. Average solubility screen values   186 T.maritima proteinsa  Rep68    Soluble  Insoluble  Mixed  Screenb  Domainc  aAverage values given for cells expressing recombinant T.maritima proteins which are either soluble, insoluble or mixed solubility.  bAverage values given for 564 clones expressing truncated Rep68 proteins.  cValues for fragment of Rep68 showing improved solubility.  Relative β-galactosidase activity  96  680  280  280  3.1  Relative Ni-HRP absorbance  780  95  530  560  1700  Solubility score  300  2.9  83  3.7  550    186 T.maritima proteinsa  Rep68    Soluble  Insoluble  Mixed  Screenb  Domainc  aAverage values given for cells expressing recombinant T.maritima proteins which are either soluble, insoluble or mixed solubility.  bAverage values given for 564 clones expressing truncated Rep68 proteins.  cValues for fragment of Rep68 showing improved solubility.  Relative β-galactosidase activity  96  680  280  280  3.1  Relative Ni-HRP absorbance  780  95  530  560  1700  Solubility score  300  2.9  83  3.7  550  View Large Fig. 1. View largeDownload slide Expression of LCK and PLA. Proteins were expressed and fractionated as described previously. Equivalent fractional volumes were visualized by 4–20% SDS–PAGE. Lane 1, MW markers; lanes 2–5, PLA fractions; lanes 6–9, LCK fractions. Whole cell extracts after induction (lanes 3 and 7) show marked overexpression of the target proteins versus the preinduction controls (lanes 2 and 6). Overexpressed PLA is found almost exclusively in the soluble fraction versus the insoluble fraction (lane 4 versus lane 5), whereas LCK is found exclusively in the insoluble fraction (lane 8 versus lane 9). Fig. 1. View largeDownload slide Expression of LCK and PLA. Proteins were expressed and fractionated as described previously. Equivalent fractional volumes were visualized by 4–20% SDS–PAGE. Lane 1, MW markers; lanes 2–5, PLA fractions; lanes 6–9, LCK fractions. Whole cell extracts after induction (lanes 3 and 7) show marked overexpression of the target proteins versus the preinduction controls (lanes 2 and 6). Overexpressed PLA is found almost exclusively in the soluble fraction versus the insoluble fraction (lane 4 versus lane 5), whereas LCK is found exclusively in the insoluble fraction (lane 8 versus lane 9). Fig. 2. View largeDownload slide Summary of screening results for 18 T.maritima proteins with pre-determined expression characteristics. (A–C) Average relative β-galactosidase activity, Ni-HRP activity and the resulting solubility scores for 18 T.maritima proteins are shown. Expression characteristics for the 18 proteins were previously determined by SDS–PAGE of both soluble and insoluble fractions. (D) SDS–PAGE of representative T.maritima proteins showing soluble and insoluble fractions of cell lysates included in average values above. Lane 1, markers (sizes in kDa); lane 2, soluble lysate fraction (TM0414); lane 3, insoluble fraction (TM0414); lane 4, soluble lysate fraction (TM0688); lane 5, insoluble fraction (TM0688); lane 6, soluble lysate fraction (TM0564); lane 7, insoluble fraction (TM0564). Fig. 2. View largeDownload slide Summary of screening results for 18 T.maritima proteins with pre-determined expression characteristics. (A–C) Average relative β-galactosidase activity, Ni-HRP activity and the resulting solubility scores for 18 T.maritima proteins are shown. Expression characteristics for the 18 proteins were previously determined by SDS–PAGE of both soluble and insoluble fractions. (D) SDS–PAGE of representative T.maritima proteins showing soluble and insoluble fractions of cell lysates included in average values above. Lane 1, markers (sizes in kDa); lane 2, soluble lysate fraction (TM0414); lane 3, insoluble fraction (TM0414); lane 4, soluble lysate fraction (TM0688); lane 5, insoluble fraction (TM0688); lane 6, soluble lysate fraction (TM0564); lane 7, insoluble fraction (TM0564). Fig. 3. View largeDownload slide Relative β-galactosidase activity versus the relative Ni-HRP activity observed after expression of 186 T.maritima proteins in reporter strain. Classification of each protein as soluble, insoluble (A) or mixed (B) is based on SDS–PAGE performed on the soluble and insoluble lysates after the screen. Fig. 3. View largeDownload slide Relative β-galactosidase activity versus the relative Ni-HRP activity observed after expression of 186 T.maritima proteins in reporter strain. Classification of each protein as soluble, insoluble (A) or mixed (B) is based on SDS–PAGE performed on the soluble and insoluble lysates after the screen. Fig. 4. View largeDownload slide Alignment of secondary structure predictions and both predicted and identified domains of Rep68. Chou–Fasman secondary structure predictions of α-helical and β-sheet structures aligned with a Kyte–Doolittle plot of hydrophilicity based on the primary sequence of Rep68. Also, aligned below are blocks representing the relative size and position of the full-length Rep68 protein, the three predicted domains of Rep68 and the Rep68 domain identified by screening of randomly generated fragments of the rep68 gene. Solubility scores for the proteins are indicated. Fig. 4. View largeDownload slide Alignment of secondary structure predictions and both predicted and identified domains of Rep68. Chou–Fasman secondary structure predictions of α-helical and β-sheet structures aligned with a Kyte–Doolittle plot of hydrophilicity based on the primary sequence of Rep68. Also, aligned below are blocks representing the relative size and position of the full-length Rep68 protein, the three predicted domains of Rep68 and the Rep68 domain identified by screening of randomly generated fragments of the rep68 gene. Solubility scores for the proteins are indicated. 4 Present address: Syrrx Corporation, 10450 Science Center Drive, Suite 100, San Diego, CA 92121, USA 2 To whom correspondence should be addressed. E-mail: lesley@gnf.org The authors would like to thank Dan Giang, Tanya Shin, Juli Vincent and Dan McMullan for their help on expression profiling and protein purification. Cloning and expression of Thermotoga proteins was supported in part by grant GM62411-02. References Allen,S.P., Polazzi,J.O., Gierse,J.K. and Easton,A.M. ( 1992) J. Bacteriol. , 174, 6938–6947. Google Scholar Altschul,S.F., Madden,T.L., Schäffer,A.A., Zhang,J., Zhang,Z., Miller,W. and Lipman,D.J. ( 1997) Nucleic Acids Res. , 25, 3389–3402. Google Scholar Bae,W., Xia,B., Inuoye,M. and Severinov,K. ( 2000) Proc. Natl Acad. Sci. USA , 97, 7784–7789. Google Scholar Beckmann,R.P., Mizzen,L.A. and Welch,W.J. ( 1990) Science , 248, 850–854. Google Scholar Caldas,T., Binet,E., Bouloc,P., Costa,A., Desgres,J. and Richarme,G. ( 2000) J. Biol. Chem. , 275, 16414–16419. Google Scholar Caldas,T., Binet,E., Bouloc,P. and Richarme,G. ( 2000) Biochem. Biophys. Res. Commun. , 271, 714–718. Google Scholar Cashel,M., Gentry,D.R., Hernandez,V.J. and Vinella,D. (1996) In Neidhardt,F.C. (ed.), Escherichia coli and Salmonella Cellular and Molecular Biology. ASM Press, Washington DC, pp. 1458–1496. Google Scholar Gething,M.-J. ( 1997) Nature  (London), 388, 329–331. Google Scholar Goloubinoff,P., Mogk,A., Zui,A.P.B., Tomoyasu,T. and Bukau,B. ( 1999) Proc. Natl Acad. Sci. USA , 96, 12732–12737. Google Scholar Hartl,F.U. ( 1996) Nature  (London), 381, 571–580. Google Scholar Jiang,W., Hou,Y. and Inouye,M. ( 1997) J. Biol. Chem. , 272, 196–202. Google Scholar Korber,P., Stahl,J.M., Nierhaus,K.H. and Bardwell,J.C.A. ( 2000) EMBO J. , 19, 741–748. Google Scholar Liberek,K. and Georgopoulos,C. ( 1993) Proc. Natl Acad. Sci. USA , 90, 11019–11023. Google Scholar Lockhart,D.J., Dong,H., Byrne,M.C., Follettie,M.T., Gallo,M.V., Chee,M.S., Mittmann,M., Wang,C., Kobayashi,M., Horton,H. and Brown,E.L. ( 1996) Nat. Biotechnol. , 14, 1675–1680. Google Scholar Maxwell,K.L., Mittermaier,A.K., Forman-Kay,J.D., and Davidson,A.R. ( 1999) Protein Sci. , 8, 1908–1911. Google Scholar McCarty,J.S., Rudiger,S., Schonfeld,H.-J., Schneider-Mergener,J., Nakahigashi,K., Yura,T. and Bukau,B. ( 1996) J. Mol. Biol. , 256, 829–837. Google Scholar Mogk,A., Tomoyasu,T., Goloubinoff,P., Rudiger,S., Roder,D. Langen,H. and Bukau,B. ( 1999) EMBO J. , 18, 6934–6949. Google Scholar Parsell,D.A. and Sauer,R.T. ( 1989) Genes Dev. , 3, 1226–1232. Google Scholar Puglisi,J.D., Blanchard,S.C. and Green,R. ( 2000) Nat. Struct. Biol. , 7, 855–861. Google Scholar Richmond,C.S., Glasner,J.D., Mau,R., Hongfan,J. and Blattner,F.R. ( 1999) Nucleic Acids Res. , 27, 3821–3835. Google Scholar Sambrook,J., Fritsch,E.F. and Maniatis,T. (1989) In Molecular Cloning: A Laboratory Manual. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, p. 1735. Google Scholar Thomas,J.G. and Baneyx,F. ( 1998) J. Bacteriol. , 180, 5165–5172. Google Scholar Veinger,L., Diamant,S., Buchner,J. and Goloubinoff,P. ( 1998) J. Biol. Chem. , 273, 11032–11037. Google Scholar Waldo,G.S., Standish,B.M., Berendzen,J. and Terwilliger,T.C. ( 1999) Nat. Biotechnol. , 17, 691–695. Google Scholar Wang,N., Yamanake,K. and Inouye,M. ( 1999) J. Bacteriol. , 181, 1603–1609. Google Scholar Wigley,W.C., Stidham,R.D., Smith,N.M., Hunt,J.F. and Thomas,P.J. ( 2001) Nat. Biotechnol. , 19, 131–135. Google Scholar Wodicka,L., Dong,H., Mittmann,M., Ho,M.H., and Lockhart,D.J. (0 1997) Nat. Biotechnol. , 15, 1359–1367. Google Scholar © Oxford University Press

Journal

Protein Engineering, Design and SelectionOxford University Press

Published: Feb 1, 2002

There are no references for this article.