Development of a novel artificial intelligence algorithm to detect pulmonary nodules on chest radiography.

BACKGROUND: In this study, we aimed to develop a novel artificial intelligence (AI) algorithm to support pulmonary nodule detection, which will enable physicians to efficiently interpret chest radiographs for lung cancer diagnosis. METHODS: We analyzed chest X-ray images obtained from a health examination center in Fukushima and the National Institutes of Health (NIH) Chest X-ray 14 dataset. We categorized these data into two types: type A included both Fukushima and NIH datasets, and type B included only the Fukushima dataset. We also demonstrated pulmonary nodules in the form of a heatmap display on each chest radiograph and calculated the positive probability score as an index value. RESULTS: Our novel AI algorithms had a receiver operating characteristic (ROC) area under the curve (AUC) of 0.74, a sensitivity of 0.75, and a specificity of 0.60 for the type A dataset. For the type B dataset, the respective values were 0.79, 0.72, and 0.74. The algorithms in both the type A and B datasets were superior to the accuracy of radiologists and similar to previous studies. CONCLUSIONS: The proprietary AI algorithms had a similar accuracy for interpreting chest radiographs when compared with previous studies and radiologists. Especially, we could train a high quality AI algorithm, even with our small type B data set. However, further studies are needed to improve and further validate the accuracy of our AI algorithm.

Streamlined human antibody generation and optimization by exploiting designed immunoglobulin loci in a B cell line.

Monoclonal antibodies (mAbs) are widely utilized as therapeutic drugs for various diseases, such as cancer, autoimmune diseases, and infectious diseases. Using the avian-derived B cell line DT40, we previously developed an antibody display technology, namely, the ADLib system, which rapidly generates antigen-specific mAbs. Here, we report the development of a human version of the ADLib system and showcase the streamlined generation and optimization of functional human mAbs. Tailored libraries were first constructed by replacing endogenous immunoglobulin genes with designed human counterparts. From these libraries, clones producing full-length human IgGs against distinct antigens can be isolated, as exemplified by the selection of antagonistic mAbs. Taking advantage of avian biology, effective affinity maturation was achieved in a straightforward manner by seamless diversification of the parental clones into secondary libraries followed by single-cell sorting, quickly affording mAbs with improved affinities and functionalities. Collectively, we demonstrate that the human ADLib system could serve as an integrative platform with unique diversity for rapid de novo generation and optimization of therapeutic or diagnostic antibody leads. Furthermore, our results suggest that libraries can be constructed by introducing exogenous genes into DT40 cells, indicating that the ADLib system has the potential to be applied for the rapid and effective directed evolution and optimization of proteins in various fields beyond biomedicine.

Crystallographic and cryogenic electron microscopic structures and enzymatic characterization of sulfur oxygenase reductase from Sulfurisphaera tokodaii.

Sulfur oxygenase reductases (SORs) are present in thermophilic and mesophilic archaea and bacteria, and catalyze oxygen-dependent oxygenation and disproportionation of elemental sulfur. SOR has a hollow, spherical homo-24-mer structure and reactions take place at active sites inside the chamber. The crystal structures of SORs from Acidianus species have been reported. However, the states of the active site components (mononuclear iron and cysteines) and the entry and exit paths of the substrate and products are still in dispute. Here, we report the biochemical and structural characterizations of SORs from the thermoacidophilic archaeon Sulfurisphaera tokodaii (StSOR) and present high-resolution structures determined by X-ray crystallography and cryogenic electron microscopy (cryo-EM). The crystal structure of StSOR was determined at 1.73 Å resolution. At the catalytic center, iron is ligated to His86, His90, Glu114, and two water molecules. Three conserved cysteines in the cavity are located 9.5-13 Å from the iron and were observed as free thiol forms. A mutational analysis indicated that the iron and one of the cysteines (Cys31) were essential for both activities. The cryo-EM structure was determined at 2.24 Å resolution using an instrument operating at 200 kV. The two structures determined by different methodologies showed similar main chain traces, but the maps exhibited different features at catalytically important components. A possible role of StSOR in the sulfur metabolism of S. tokodaii (an obligate aerobe) is discussed based on this study. Given the high resolution achieved in this study, StSOR was shown to be a good benchmark sample for cryo-EM.

Cold shock proteins improve E. coli cell-free synthesis in terms of soluble yields of aggregation-prone proteins.

Protein folding is usually slowed-down at low temperatures, and thus low-temperature expression is an effective strategy to improve the soluble yield of aggregation-prone proteins. In this study, we investigated the effects of a variety of cold shock proteins and domains (Csps) on an Escherichia coli cell extract-based cell-free protein synthesis system (CF). Most of the 12 Csps that were successfully prepared dramatically improved the protein yields, by factors of more than 5 at 16°C and 2 at 23°C, to levels comparable to those obtained at 30°C. Their stimulatory effects were complementary to each other, while CspD and CspH were inhibitory. The Csps' effects correlated well with their Pfam CSD family scores (PF00313.22). All of the investigated Csps, except CspH, similarly possessed RNA binding and chaperon activities and increased the messenger RNA amount irrespective of their effect, suggesting that the proper balance between these activities was required for the enhancement. Unexpectedly, the 5'-untranslated region of cspA was less effective as the leader sequence. Our results demonstrated that the use of the Csps presented in this study will provide a simple and highly effective strategy for the CF, to improve the soluble yields of aggregation-prone proteins.

Anti-Semaphorin 3A neutralization monoclonal antibody prevents sepsis development in lipopolysaccharide-treated mice.

Semaphorin 3A (Sema3A), originally identified as a potent growth cone collapsing factor in developing sensory neurons, is now recognized as a key player in immune, cardiovascular, bone metabolism and neurological systems. Here we established an anti-Sema3A monoclonal antibody that neutralizes the effects of Sema3A both in vitro and in vivo. The anti-Sema3A neutralization chick IgM antibodies were screened by combining an autonomously diversifying library selection system and an in vitro growth cone collapse assay. We further developed function-blocking chick-mouse chimeric and humanized anti-Sema3A antibodies. We found that our anti-Sema3A antibodies were effective for improving the survival rate in lipopolysaccharide-induced sepsis in mice. Our antibody is a potential therapeutic agent that may prevent the onset of or alleviate symptoms of human diseases associated with Sema3A.

Sorting nexin 27 interacts with multidrug resistance-associated protein 4 (MRP4) and mediates internalization of MRP4.

Multidrug resistance-associated protein 4 (MRP4/ABCC4) makes a vital contribution to the bodily distribution of drugs and endogenous compounds because of its cellular efflux abilities. However, little is known about the mechanism regulating its cell surface expression. MRP4 has a PDZ-binding motif, which is a potential sequence that modulates the membrane expression of MRP4 via interaction with PDZ adaptor proteins. To investigate this possible relationship, we performed GST pull-down assays and subsequent analysis with matrix-assisted laser desorption/ionization-time of flight mass spectrometry. This method identified sorting nexin 27 (SNX27) as the interacting PDZ adaptor protein with a PDZ-binding motif of MRP4. Its interaction was confirmed by a coimmunoprecipitation study using HEK293 cells. Knockdown of SNX27 by siRNA in HEK293 cells raised MRP4 expression on the plasma membrane, increased the extrusion of 6-[(14)C]mercaptopurine, an MRP4 substrate, and conferred resistance against 6-[(14)C]mercaptopurine. Cell surface biotinylation studies indicated that the inhibition of MRP4 internalization was responsible for these results. Immunocytochemistry and cell surface biotinylation studies using COS-1 cells showed that SNX27 localized to both the early endosome and the plasma membrane. These data suggest that SNX27 interacts with MRP4 near the plasma membrane and promotes endocytosis of MRP4 and thereby negatively regulates its cell surface expression and transport function.

AP2 adaptor complex mediates bile salt export pump internalization and modulates its hepatocanalicular expression and transport function.

UNLABELLED: The bile salt export pump (BSEP) mediates the biliary excretion of bile salts and its dysfunction induces intrahepatic cholestasis. Reduced canalicular expression of BSEP resulting from the promotion of its internalization is one of the causes of this disease state. However, the molecular mechanism underlying BSEP internalization from the canalicular membrane (CM) remains unknown. We have shown previously that 4-phenylbutyrate (4PBA), a drug used for ornithine transcarbamylase deficiency (OTCD), inhibited internalization and subsequent degradation of cell-surface-resident BSEP. The current study found that 4PBA treatment decreased significantly the expression of α- and µ2-adaptin, both of which are subunits of the AP2 adaptor complex (AP2) that mediates clathrin-dependent endocytosis, in liver specimens from rats and patients with OTCD, and that BSEP has potential AP2 recognition motifs in its cytosolic region. Based on this, the role of AP2 in BSEP internalization was explored further. In vitro analysis with 3×FLAG-human BSEP-expressing HeLa cells and human sandwich-culture hepatocytes indicates that the impairment of AP2 function by RNA interference targeting of α-adaptin inhibits BSEP internalization from the plasma membrane and increases its cell-surface expression and transport function. Studies using immunostaining, coimmunoprecipitation, glutathione S-transferase pulldown assay, and time-lapse imaging show that AP2 interacts with BSEP at the CM through a tyrosine motif at the carboxyl terminus of BSEP and mediates BSEP internalization from the CM of hepatocytes. CONCLUSION: AP2 mediates the internalization and subsequent degradation of CM-resident BSEP through direct interaction with BSEP and thereby modulates the canalicular expression and transport function of BSEP. This information should be useful for understanding the pathogenesis of severe liver diseases associated with intrahepatic cholestasis.

4-Phenylbutyrate modulates ubiquitination of hepatocanalicular MRP2 and reduces serum total bilirubin concentration.

BACKGROUND & AIMS: Multidrug resistance-associated protein 2 (in humans, MRP2; in rodents, Mrp2) mediates biliary excretion of bilirubin glucuronides. Therefore, upregulation of MRP2/Mrp2 expression may improve hyperbilirubinemia. We investigated the effects of 4-phenylbutyrate (4PBA), a drug used to treat ornithine transcarbamylase deficiency (OTCD), on the cell surface expression and transport function of MRP2/Mrp2 and serum T-Bil concentration. METHODS: MRP2-expressing MDCKII (MRP2-MDCKII) cells and rats were studied to explore the change induced by 4PBA treatment in the cell surface expression and transport function of MRP2/Mrp2 and its underlying mechanism. Serum and liver specimens from OTCD patients were analyzed to examine the effect of 4PBA on hepatic MRP2 expression and serum T-Bil concentration in humans. RESULTS: In MRP2-MDCKII cells and the rat liver, 4PBA increased the cell surface expression and transport function of MRP2/Mrp2. In patients with OTCD, hepatic MRP2 expression increased and serum T-Bil concentration decreased significantly after 4PBA treatment. In vitro studies designed to explore the mechanism underlying this drug action suggested that cell surface-resident MRP2/Mrp2 is degraded via ubiquitination-mediated targeting to the endosomal/lysosomal degradation pathway and that 4PBA inhibits the degradation of cell surface-resident MRP2/Mrp2 by reducing its susceptibility to ubiquitination. CONCLUSIONS: 4PBA activates MRP2/Mrp2 function through increased expression of MRP2/Mrp2 at the hepatocanalicular membrane by modulating its ubiquitination, and thereby decreases serum T-Bil concentration. 4PBA has thus therapeutic potential for improving hyperbilirubinemia.

Structural insight into the zinc finger CW domain as a histone modification reader.

The zinc finger CW (zf-CW) domain is a motif of about 60 residues that is frequently found in proteins involved in epigenetic regulation. Here, we determined the NMR solution structure of the zf-CW domain of the human zf-CW and PWWP domain containing protein 1 (ZCWPW1). The zf-CW domain adopts a new fold in which a zinc ion is coordinated tetrahedrally by four conserved Cys ligand residues. The tertiary structure of the zf-CW domain partially resembles that adopted by the plant homeo domain (PHD) finger bound to the histone tail, suggesting that the zf-CW domain and the PHD finger have similar functions. The solution structure of the complex of the zf-CW domain with the histone H3 tail peptide (1-10) with trimethylated K4 clarified its binding mode. Our structural and biochemical studies have identified the zf-CW domain as a member of the histone modification reader modules for epigenetic regulation.

Structural basis for the recognition of nucleophosmin-anaplastic lymphoma kinase oncoprotein by the phosphotyrosine binding domain of Suc1-associated neurotrophic factor-induced tyrosine-phosphorylated target-2.

The nucleophosmin-anaplastic lymphoma kinase (NPM-ALK) fusion oncoprotein, formed by the t(2;5) chromosomal translocation in anaplastic large-cell lymphomas, has constitutive tyrosine kinase activity and interacts with a number of signaling molecules. One of the interacting partners of NPM-ALK is the adaptor protein, Suc1-associated neurotrophic factor-induced tyrosine-phosphorylated target (SNT), and mutations that deprive NPM-ALK of all three of the SNT-binding sites significantly reduced the transforming activity. In this study, the interactions of the three binding sites in NPM-ALK with the phosphotyrosine binding (PTB) domain of SNT-2 were analyzed. First, by isothermal titration calorimetry, we found that the phosphorylation-independent binding site in NPM-ALK interacts with the SNT-2 PTB domain more tightly than the phosphorylation-dependent binding sites. Second, the solution structure of the SNT-2 PTB domain in complex with the nonphosphorylated NPM-ALK peptide was determined by nuclear magnetic resonance spectroscopy. The NPM-ALK peptide interacts with the hydrophobic surface of the PTB domain and intermolecularly extends the PTB beta-sheet. This interaction mode is much broader and more extensive than those of the phosphorylation-dependent binding sites. Our results indicate that the higher binding activity of the phosphorylation-independent binding site is caused by additional hydrophobic interactions.

Site-specific incorporation of 4-iodo-L-phenylalanine through opal suppression.

A variety of unique codons have been employed to expand the genetic code. The use of the opal (UGA) codon is promising, but insufficient information is available about the UGA suppression approach, which facilitates the incorporation of non-natural amino acids through suppression of the UGA codon. In this study, the UGA codon was used to incorporate 4-iodo-l-phenylalanine into position 32 of the Ras protein in an Escherichia coli cell-free translation system. The undesired incorporation of tryptophan in response to the UGA codon was completely repressed by the addition of indolmycin. The minor amount (3%) of contaminating 4-bromo-l-phenylalanine in the building block 4-iodo-l-phenylalanine led to the significant incorporation of 4-bromo-l-phenylalanine (21%), and this problem was solved by using a purified 4-iodo-l-phenylalanine sample. Optimization of the incubation time was also important, since the undesired incorporation of free phenylalanine increased during the cell-free translation reaction. The 4-iodo-l-phenylalanine residue can be used for the chemoselective modification of proteins. This method will contribute to advancements in protein engineering studies with non-natural amino acid substitutions.

Mathematical model for empirically optimizing large scale production of soluble protein domains.

BACKGROUND: Efficient dissection of large proteins into their structural domains is critical for high throughput proteome analysis. So far, no study has focused on mathematically modeling a protein dissection protocol in terms of a production system. Here, we report a mathematical model for empirically optimizing the cost of large-scale domain production in proteomics research. RESULTS: The model computes the expected number of successfully producing soluble domains, using a conditional probability between domain and boundary identification. Typical values for the model's parameters were estimated using the experimental results for identifying soluble domains from the 2,032 Kazusa HUGE protein sequences. Among the 215 fragments corresponding to the 24 domains that were expressed correctly, 111, corresponding to 18 domains, were soluble. Our model indicates that, under the conditions used in our pilot experiment, the probability of correctly predicting the existence of a domain was 81% (175/215) and that of predicting its boundary was 63% (111/175). Under these conditions, the most cost/effort-effective production of soluble domains was to prepare one to seven fragments per predicted domain. CONCLUSIONS: Our mathematical modeling of protein dissection protocols indicates that the optimum number of fragments tested per domain is actually much smaller than expected a priori. The application range of our model is not limited to protein dissection, and it can be utilized for designing various large-scale mutational analyses or screening libraries.

Solution structure of the RNA binding domain in the human muscleblind-like protein 2.

The muscleblind-like (MBNL) proteins 1, 2, and 3, which contain four CCCH zinc finger motifs (ZF1-4), are involved in the differentiation of muscle inclusion by controlling the splicing patterns of several pre-mRNAs. Especially, MBNL1 plays a crucial role in myotonic dystrophy. The CCCH zinc finger is a sequence motif found in many RNA binding proteins and is suggested to play an important role in the recognition of RNA molecules. Here, we solved the solution structures of both tandem zinc finger (TZF) motifs, TZF12 (comprising ZF1 and ZF2) and TZF34 (ZF3 and ZF4), in MBNL2 from Homo sapiens. In TZF12 of MBNL2, ZF1 and ZF2 adopt a similar fold, as reported previously for the CCCH-type zinc fingers in the TIS11d protein. The linker between ZF1 and ZF2 in MBNL2 forms an antiparallel beta-sheet with the N-terminal extension of ZF1. Furthermore, ZF1 and ZF2 in MBNL2 interact with each other through hydrophobic interactions. Consequently, TZF12 forms a single, compact global fold, where ZF1 and ZF2 are approximately symmetrical about the C2 axis. The structure of the second tandem zinc finger (TZF34) in MBNL2 is similar to that of TZF12. This novel three-dimensional structure of the TZF domains in MBNL2 provides a basis for functional studies of the CCCH-type zinc finger motifs in the MBNL protein family.

Structure of the C-terminal phosphotyrosine interaction domain of Fe65L1 complexed with the cytoplasmic tail of amyloid precursor protein reveals a novel peptide binding mode.

Fe65L1, a member of the Fe65 family, is an adaptor protein that interacts with the cytoplasmic domain of Alzheimer amyloid precursor protein (APP) through its C-terminal phosphotyrosine interaction/phosphotyrosine binding (PID/PTB) domain. In the present study, the solution structures of the C-terminal PID domain of mouse Fe65L1, alone and in complex with a 32-mer peptide (DAAVTPEERHLSKMQQNGYENPTYKFFEQMQN) derived from the cytoplasmic domain of APP, were determined using NMR spectroscopy. The C-terminal PID domain of Fe65L1 alone exhibits a canonical PID/PTB fold, whereas the complex structure reveals a novel mode of peptide binding. In the complex structure, the NPTY motif forms a type-I beta-turn, and the residues immediately N-terminal to the NPTY motif form an antiparallel beta-sheet with the beta5 strand of the PID domain, the binding mode typically observed in the PID/PTB.peptide complex. On the other hand, the N-terminal region of the peptide forms a 2.5-turn alpha-helix and interacts extensively with the C-terminal alpha-helix and the peripheral regions of the PID domain, representing a novel mode of peptide binding that has not been reported previously for the PID/PTB.peptide complex. The indispensability of the N-terminal region of the peptide for the high affinity of the PID-peptide interaction is consistent with NMR titration and isothermal calorimetry data. The extensive binding features of the PID domain of Fe65L1 with the cytoplasmic domain of APP provide a framework for further understanding of the function, trafficking, and processing of APP modulated by adapter proteins.

Solution structure of the zinc finger HIT domain in protein FON.

The zinc finger HIT domain is a sequence motif found in many proteins, including thyroid hormone receptor interacting protein 3 (TRIP-3), which is possibly involved in maturity-onset diabetes of the young (MODY). Novel zinc finger motifs are suggested to play important roles in gene regulation and chromatin remodeling. Here, we determined the high-resolution solution structure of the zinc finger HIT domain in ZNHIT2 (protein FON) from Homo sapiens, by an NMR method based on 567 upper distance limits derived from NOE intensities measured in three-dimensional NOESY spectra. The structure yielded a backbone RMSD to the mean coordinates of 0.19 A for the structured residues 12-48. The fold consists of two consecutive antiparallel beta-sheets and two short C-terminal helices packed against the second beta-sheet, and binds two zinc ions. Both zinc ions are coordinated tetrahedrally via a CCCC-CCHC motif to the ligand residues of the zf-HIT domain in an interleaved manner. The tertiary structure of the zinc finger HIT domain closely resembles the folds of the B-box, RING finger, and PHD domains with a cross-brace zinc coordination mode, but is distinct from them. The unique three-dimensional structure of the zinc finger HIT domain revealed a novel zinc-binding fold, as a new member of the treble clef domain family. On the basis of the structural data, we discuss the possible functional roles of the zinc finger HIT domain.

Solution structure of the general transcription factor 2I domain in mouse TFII-I protein.

The general transcription factor TFII-I, with the corresponding gene name GTF2I, is an unusual transcriptional regulator that associates with both basal and signal-induced transcription factors. TFII-I consists of six GTF2I repeat domains, called I-repeats R1-R6. The structure and function of the GTF2I domain are not clearly understood, even though it contains a helix-loop-helix motif, which is considered to be the protein-protein interaction area, based on biochemical analyses. Here, we report the solution structure of the fifth repeat of the six GTF2I repeat domains from murine TFII-I, which was determined by heteronuclear multidimensional NMR spectroscopy (PDB code 1Q60). The three-dimensional structure of the GTF2I domain is classified as a new fold, consisting of four helices (residues 8-24, 34-39, 63-71, and 83-91), two antiparallel beta strands (residues 44-47 and 77-80), and a well-defined loop containing two beta-turns between sheet 1 and helix 3. All of the repeats probably have similar folds to that of repeat 5, because the conserved residues in the GTF2I repeat domains are assembled on the hydrophobic core, turns, and secondary structure elements, as revealed by a comparison of the sequences of the first through the sixth GTF2I repeats in TFII-I.

Structural and functional differences of SWIRM domain subtypes.

SWIRM is a conserved domain found in several chromatin-associated proteins. Based on their sequences, the SWIRM family members can be classified into three subfamilies, which are represented by Swi3, LSD1, and Ada2. Here we report the SWIRM structure of human MYb-like, Swirm and Mpn domain-containing protein-1 (MYSM1). The MYSM1 SWIRM structure forms a compact HTH-related fold comprising five alpha-helices, which best resembles the Swi3 SWIRM structure, among the known SWIRM structures. The MYSM1 and Swi3 SWIRM structures are more similar to the LSD1 structure than the Ada2alpha structure. The SWIRM domains of MYSM1 and LSD1 lacked DNA binding activity, while those of Ada2alpha and the human Swi3 counterpart, SMARCC2, bound DNA. The dissimilarity in the DNA-binding ability of the MYSM1 and SMARCC2 SWIRM domains might be due to a couple of amino acid differences in the last helix. These results indicate that the SWIRM family has indeed diverged into three structural subfamilies (Swi3/MYSM1, LSD1, and Ada2 types), and that the Swi3/MYSM1-type subfamily has further diverged into two functionally distinct groups. We also solved the structure of the SANT domain of MYSM1, and demonstrated that it bound DNA with a similar mode to that of the c-Myb DNA-binding domain.

Improving cell-free protein synthesis for stable-isotope labeling.

Cell-free protein synthesis is suitable for stable-isotope labeling of proteins for NMR analysis. The Escherichia coli cell-free system containing potassium acetate for efficient translation (KOAc system) is usually used for stable-isotope labeling, although it is less productive than other systems. A system containing a high concentration of potassium L-glutamate (L-Glu system), instead of potassium acetate, is highly productive, but cannot be used for stable-isotope labeling of Glu residues. In this study, we have developed a new cell-free system that uses potassium D-glutamate (D-Glu system). The productivity of the D-Glu system is approximately twice that of the KOAc system. The cross peak intensities in the 1H-15N HSQC spectrum of the uniformly stable-isotope labeled Ras protein, prepared with the D-Glu system, were similar to those obtained with the KOAc system, except that the Asp intensities were much higher for the protein produced with the D-Glu system. These results indicate that the D-Glu system is a highly productive cell-free system that is especially useful for stable-isotope labeling of proteins.

Site-specific functionalization of proteins by organopalladium reactions.

A new carbon-carbon bond has been regioselectively introduced into a target position (position 32 or 174) of the Ras protein by two types of organopalladium reactions (Mizoroki-Heck and Sonogashira reactions). Reaction conditions were screened by using a model peptide, and the stability of the Ras protein under the reaction conditions was examined by using the wild-type Ras protein. Finally, the iF-Ras proteins containing a 4-iodo-L-phenylalanine residue were subjected to organopalladium reactions with vinylated or propargylated biotin. Site-specific biotinylations of the Ras protein were confirmed by Western blot and LC-MS/MS.

A robust two-step PCR method of template DNA production for high-throughput cell-free protein synthesis.

A two-step PCR method has been developed for the robust, high-throughput production of linear templates ready for cell-free protein synthesis. The construct made from the cDNA expresses a target protein region with N- and/or C-terminal tags. The procedure consists only of mixing, dilution, and PCR steps, and is free from cloning and purification steps. In the first step of the two-step PCR, a target region within the coding sequence is amplified using two gene-specific forward and reverse primers, which contain the linker sequences and the terminal sequences of the target region. The second PCR concatenates the first PCR product with the N- and C-terminal double-stranded fragments, which contain the linker sequences as well as the sequences for the tag(s) and the initiation and termination, respectively, for T7 transcription and ribosomal translation, and amplifies it with the universal primer. Proteins can be fused with a variety of tags, such as natural poly-histidine, glutathione-S-transferase, maltose-binding protein, and/or streptavidin-binding peptide. The two-step PCR method was successfully applied to 42 human target protein regions with various GC contents (38-77%). The robustness of the two-step PCR method against possible fluctuations of experimental conditions in practical use was explored. The second PCR product was obtained at 60-120 microg/ml, and was used without purification as a template at a concentration of 2-4 microg/ml in an Escherichia coli coupled transcription-translation system. This combination of two-step PCR with cell-free protein synthesis is suitable for the rapid production of proteins in milligram quantities for genome-scale studies.