Direct epitranscriptomic regulation of mammalian translation initiation through N4-acetylcytidine.

mRNA function is influenced by modifications that modulate canonical nucleobase behavior. We show that a single modification mediates distinct impacts on mRNA translation in a position-dependent manner. Although cytidine acetylation (ac4C) within protein-coding sequences stimulates translation, ac4C within 5' UTRs impacts protein synthesis at the level of initiation. 5' UTR acetylation promotes initiation at upstream sequences, competitively inhibiting annotated start codons. Acetylation further directly impedes initiation at optimal AUG contexts: ac4C within AUG-flanking Kozak sequences reduced initiation in base-resolved transcriptome-wide HeLa results and in vitro utilizing substrates with site-specific ac4C incorporation. Cryo-EM of mammalian 80S initiation complexes revealed that ac4C in the -1 position adjacent to an AUG start codon disrupts an interaction between C and hypermodified t6A at nucleotide 37 of the initiator tRNA. These findings demonstrate the impact of RNA modifications on nucleobase function at a molecular level and introduce mRNA acetylation as a factor regulating translation in a location-specific manner.

A Noncoding A-to-U Kozak Site Change Related to the High Transmissibility of Alpha, Delta, and Omicron VOCs.

Three prevalent SARS-CoV-2 variants of concern (VOCs) emerged and caused epidemic waves. It is essential to uncover advantageous mutations that cause the high transmissibility of VOCs. However, viral mutations are tightly linked, so traditional population genetic methods, including machine learning-based methods, cannot reliably detect mutations conferring a fitness advantage. In this study, we developed an approach based on the sequential occurrence order of mutations and the accelerated furcation rate in the pandemic-scale phylogenomic tree. We analyzed 3,777,753 high-quality SARS-CoV-2 genomic sequences and the epidemiology metadata using the Coronavirus GenBrowser. We found that two noncoding mutations at the same position (g.a28271-/u) may be crucial to the high transmissibility of Alpha, Delta, and Omicron VOCs although the noncoding mutations alone cannot increase viral transmissibility. Both mutations cause an A-to-U change at the core position -3 of the Kozak sequence of the N gene and significantly reduce the protein expression ratio of ORF9b to N. Using a convergent evolutionary analysis, we found that g.a28271-/u, S:p.P681H/R, and N:p.R203K/M occur independently on three VOC lineages, suggesting that coordinated changes of S, N, and ORF9b proteins are crucial to high viral transmissibility. Our results provide new insights into high viral transmissibility co-modulated by advantageous noncoding and nonsynonymous changes.

Unheeded SARS-CoV-2 proteins? A deep look into negative-sense RNA.

SARS-CoV-2 is a novel positive-sense single-stranded RNA virus from the Coronaviridae family (genus Betacoronavirus), which has been established as causing the COVID-19 pandemic. The genome of SARS-CoV-2 is one of the largest among known RNA viruses, comprising of at least 26 known protein-coding loci. Studies thus far have outlined the coding capacity of the positive-sense strand of the SARS-CoV-2 genome, which can be used directly for protein translation. However, it has been recently shown that transcribed negative-sense viral RNA intermediates that arise during viral genome replication from positive-sense viruses can also code for proteins. No studies have yet explored the potential for negative-sense SARS-CoV-2 RNA intermediates to contain protein-coding loci. Thus, using sequence and structure-based bioinformatics methodologies, we have investigated the presence and validity of putative negative-sense ORFs (nsORFs) in the SARS-CoV-2 genome. Nine nsORFs were discovered to contain strong eukaryotic translation initiation signals and high codon adaptability scores, and several of the nsORFs were predicted to interact with RNA-binding proteins. Evolutionary conservation analyses indicated that some of the nsORFs are deeply conserved among related coronaviruses. Three-dimensional protein modeling revealed the presence of higher order folding among all putative SARS-CoV-2 nsORFs, and subsequent structural mimicry analyses suggest similarity of the nsORFs to DNA/RNA-binding proteins and proteins involved in immune signaling pathways. Altogether, these results suggest the potential existence of still undescribed SARS-CoV-2 proteins, which may play an important role in the viral lifecycle and COVID-19 pathogenesis.

Conservation and Variability of the AUG Initiation Codon Context in Eukaryotes.

Selection of the translation initiation site (TIS) is a crucial step during translation. In the 1980s Marylin Kozak performed key studies on vertebrate mRNAs to characterize the optimal TIS consensus sequence, the Kozak motif. Within this motif, conservation of nucleotides in crucial positions, namely a purine at -3 and a G at +4 (where the A of the AUG is numbered +1), is essential for TIS recognition. Ever since its characterization the Kozak motif has been regarded as the optimal sequence to initiate translation in all eukaryotes. We revisit here published in silico data on TIS consensus sequences, as well as experimental studies from diverse eukaryotic lineages, and propose that, while the -3A/G position is universally conserved, the remaining variability of the consensus sequences enables their classification as optimal, strong, and moderate TIS sequences.

Kozak Sequence Acts as a Negative Regulator for De Novo Transcription Initiation of Newborn Coding Sequences in the Plant Genome.

The manner in which newborn coding sequences and their transcriptional competency emerge during the process of gene evolution remains unclear. Here, we experimentally simulated eukaryotic gene origination processes by mimicking horizontal gene transfer events in the plant genome. We mapped the precise position of the transcription start sites (TSSs) of hundreds of newly introduced promoterless firefly luciferase (LUC) coding sequences in the genome of Arabidopsis thaliana cultured cells. The systematic characterization of the LUC-TSSs revealed that 80% of them occurred under the influence of endogenous promoters, while the remainder underwent de novo activation in the intergenic regions, starting from pyrimidine-purine dinucleotides. These de novo TSSs obeyed unexpected rules; they predominantly occurred ∼100 bp upstream of the LUC inserts and did not overlap with Kozak-containing putative open reading frames (ORFs). These features were the output of the immediate responses to the sequence insertions, rather than a bias in the screening of the LUC gene function. Regarding the wild-type genic TSSs, they appeared to have evolved to lack any ORFs in their vicinities. Therefore, the repulsion by the de novo TSSs of Kozak-containing ORFs described above might be the first selection gate for the occurrence and evolution of TSSs in the plant genome. Based on these results, we characterized the de novo type of TSS identified in the plant genome and discuss its significance in genome evolution.

Fine-tuning the expression of pathway gene in yeast using a regulatory library formed by fusing a synthetic minimal promoter with different Kozak variants.

BACKGROUND: Tailoring gene expression to balance metabolic fluxes is critical for the overproduction of metabolites in yeast hosts, and its implementation requires coordinated regulation at both transcriptional and translational levels. Although synthetic minimal yeast promoters have shown many advantages compared to natural promoters, their transcriptional strength is still limited, which restricts their applications in pathway engineering. RESULTS: In this work, we sought to expand the application scope of synthetic minimal yeast promoters by enhancing the corresponding translation levels using specific Kozak sequence variants. Firstly, we chose the reported UASF-E-C-Core1 minimal promoter as a library template and determined its Kozak motif (K0). Next, we randomly mutated the K0 to generate a chimeric promoter library, which was able to drive green fluorescent protein (GFP) expression with translational strengths spanning a 500-fold range. A total of 14 chimeric promoters showed at least two-fold differences in GFP expression strength compared to the K0 control. The best one named K528 even showed 8.5- and 3.3-fold increases in fluorescence intensity compared with UASF-E-C-Core1 and the strong native constitutive promoter PTDH3, respectively. Subsequently, we chose three representative strong chimeric promoters (K540, K536, and K528) from this library to regulate pathway gene expression. In conjunction with the tHMG1 gene for squalene production, the K528 variant produced the best squalene titer of 32.1 mg/L in shake flasks, which represents a more than 10-fold increase compared to the parental K0 control (3.1 mg/L). CONCLUSIONS: All these results demonstrate that this chimeric promoter library developed in this study is an effective tool for pathway engineering in yeast.

Tailoring translational strength using Kozak sequence variants improves bispecific antibody assembly and reduces product-related impurities in CHO cells.

Optimal production of bispecific antibodies (bsAb) requires efficient and tailored co-expression and assembly of two distinct heavy and two distinct light chains. Here, we describe a novel technology to modulate the translational strength of antibody chains via Kozak sequence variants to produce bsAb in a single cell line. In this study, we designed and screened a large Kozak sequence library to identify 10 independent variants that can modulate protein expression levels from approximately 0.2 to 1.3-fold compared with the wild-type sequence in transient transfection. We used a combination of several of these variants, covering a wide range of translational strength, to develop stable single cell Chinese hamster ovary bispecific cell lines and compared the results with those obtained from the wild-type sequence. A significant increase in bispecific antibody assembly with a concomitant reduction in the level of product-related impurities was observed. Our findings suggest that for production of bsAb it can be advantageous to modify translational strength for selected protein chains to improve overall yield and product quality. By extension, tuning of translational strength can also be applied to improving the production of a wide variety of heterologous proteins.

A Mechanosensitive Channel, Mouse Transmembrane Channel-Like Protein 1 (Mtmc1) Is Translated from a Splice Variant mTmc1ex1 but Not from the Other Variant mTmc1ex2.

Mechanical stimuli caused by sound waves are detected by hair cells in the cochlea through the opening of mechanoelectrical transduction (MET) channels. Transmembrane channel-like protein 1 (TMC1) has been revealed to be the pore-forming component of the MET channel. The two splice variants for mouse Tmc1 (mTmc1ex1 and mTmc1ex2) were reported to be expressed in the cochlea of infant mice, though only the sequence of mTmc1ex2 had been deposited in GenBank. However, due to the presence of an upstream open reading frame (uORF) and the absence of a typical Kozak sequence in mTmc1ex2, we questioned whether mTMC1 was translated from mTmc1ex2. Therefore, in this study, we evaluated which splice variant was protein-coding mRNA. Firstly, the results of RT-PCR and cDNA cloning of mTmc1 using mRNA isolated from the cochlea of five-week-old mice suggested that more Tmc1ex1 were expressed than mTmc1ex2. Secondly, mTMC1 was translated from mTmc1ex1 but not from mTmc1ex2 in a heterologous expression system. Finally, analyses using site-directed mutagenesis revealed that the uORF and the weak Kozak sequence in mTmc1ex2 prevented the translation of mTMC1 from mTmc1ex2. These results suggest that mTmc1ex1 plays a main role in the expression of mTMC1 in the mouse cochlea, and therefore, mTmc1ex1 should be the mRNA for mTMC1 hereafter.

Optimization of a 2A self-cleaving peptide-based multigene expression system for efficient expression of upstream and downstream genes in silkworm.

The multigene expression system is highly attractive to co-express multiple genes or multi-subunit complex-based genes for their functional studies, and in gene therapy and visual tracking of expressed proteins. However, the current multiple gene co-expression strategies usually suffer from severe inefficiency and unbalanced expression of multiple genes. Here, we report on an improved 2A self-cleaving peptide (2A)-based multigene expression system (2A-MGES), by introducing an optimized Kozak region (Ck) and altering the gene arrangement, both of which contributed to the efficient expression of two fluorescent protein genes in silkworm. By co-expressing DsRed and EGFP genes in insect cells and silkworms, the potent Ck was first found to improve the translation efficiency of downstream genes, and the expression of the flanking genes of 2A were improved by altering the gene arrangement in 2A-MGES. Moreover, we showed that combining Ck and an optimized gene arrangement in 2A-MGES could synergistically improve the expression of genes in the cell. Further, these two flanking genes, regulated by modified 2A-MGES, were further co-expressed in the middle silk gland and secreted into the cocoon, and both achieved efficient expression in the transgenic silkworms and their cocoons. These results suggested that the modified Ck-2A-MGES will be a potent tool for multiple gene expression, for studies of their functions, and their applications in insect species.

A mild case of molybdenum cofactor deficiency defines an alternative route of MOCS1 protein maturation.

Molybdenum cofactor deficiency is an autosomal recessive inborn error of metabolism, which results from mutations in genes involved in Moco biosynthesis. Moco serves as a cofactor of several enzymes, including sulfite oxidase. MoCD is clinically characterized by intractable seizures and severe, rapidly progressing neurodegeneration leading to death in early childhood in the majority of known cases. Here we report a patient with an unusual late disease onset and mild phenotype, characterized by a lack of seizures, normal early development, a decline triggered by febrile illness and a subsequent dystonic movement disorder. Genetic analysis revealed a homozygous c.1338delG MOCS1 mutation causing a frameshift (p.S442fs) with a premature termination of the MOCS1AB translation product at position 477 lacking the entire MOCS1B domain. Surprisingly, urine analysis detected trace amounts (1% of control) of the Moco degradation product urothione, suggesting a residual Moco synthesis in the patient, which was consistent with the mild clinical presentation. Therefore, we performed bioinformatic analysis of the patient's mutated MOCS1 transcript and found a potential Kozak-sequence downstream of the mutation site providing the possibility of an independent expression of a MOCS1B protein. Following the expression of the patient's MOCS1 cDNA in HEK293 cells we detected two proteins: a truncated MOCS1AB protein and a 22.4 kDa protein representing MOCS1B. Functional studies of both proteins confirmed activity of MOCS1B, but not of the truncated MOCS1AB. This finding demonstrates an unusual mechanism of translation re-initiation in the MOCS1 transcript, which results in trace amounts of functional MOCS1B protein being sufficient to partially protect the patient from the most severe symptoms of MoCD.

In vitro and computational studies on the effects of ARE deletion and targeted mutations on the expression of interferon beta-1a in HEK293T cells.

Interferon beta (IFNß) is transiently expressed in response to viral infections and widely used to treat relapsing-remitting multiple sclerosis (MS). We introduced mutations in the IFNß gene (in the 27th and 101st codons and in the Kozak sequence, and also deletion of 3' and 5' unstable, untranslated region, UTR) with the aim of increasing the expression of IFNß. Computational analyses of mutant and wild-type RNAs and proteins of IFNß by RNAfold, ASAView, HOPE and Ramachandran plot, and iStable web servers showed that the mutations could decrease RNA stability, protein solvent accessibility, and protein stability but could not change correct folding. Two recombinant IFNß101 and IFNß101+27 constructs were designed by site-directed mutagenesis. The wild-type IFNß gene also was used as a control. In vitro experiments by quantitative real-time PCR, dot blot, SDS-PAGE, and Western blot assays showed an increased expression level of recombinant IFNßs. 79.9-fold, 99.7-fold, and 190-fold elevations in the mRNA expression of IFNßw, IFNß101, and IFNß101+27 were seen, respectively, in comparison with the endogenous IFNß mRNA in untransfected HEK293T cells. The IFNß mRNA expression was increased 2.38-fold and 1.25-fold for 101+27 and 101 mutated forms, respectively, in comparison with the IFNß wild-type construct. An elevation in IFNß protein production was also clearly detected in the transfected HEK293T cell containing recombinant IFNß101 and IFNß101+27 constructs. Finally, these directed mutations in the IFNß gene successfully elevated protein and mRNA production but in silico analyses of mutant mRNAs showed decreased mRNA stability, unlike previous studies, in comparison with the wild-type mRNA.

Identification and characterization of pyrokinin and CAPA peptides, and corresponding GPCRs from spotted wing drosophila, Drosophila suzukii.

The family of FXPRLamide peptides serves as a major insect hormone. It is characterized by a core active amino acid sequence conserved at the C-terminal ends, and provides various physiological roles across the Insecta. In this study we identified and characterized pyrokinin (PK) and CAPA cDNAs encoding two FXPRLamide peptides, pyrokinin and CAPA-DH (diapause hormone), and two corresponding G protein-coupled receptors (GPCRs) from spotted wing drosophila (SWD), Drosophila suzukii. Expressions of PK and CAPA mRNAs were differentially observed during all life stages except the embryo, and the detection of CAPA transcription was relatively strong compared with the PK gene in SWD. Both D. suzukii pyrokinin receptor (DrosuPKr) and CAPA-DH receptor (DrosuCAPA-DHr) were functionally expressed and confirmed through binding to PK and DH peptides. Differential expression of two GPCRs occurred during all life stages; a strong transcription of DrosuPKr was observed in the 3rd instar. DrosuCAPA-DHr was clearly expressed from the embryo to the larva, but not detected in the adult. Gene regulation during the life stages was not synchronized between ligand and receptor. For example, SWD CAPA mRNA has been up-regulated in the adult while CAPA-DHr was down-regulated. The difference could be from the CAPA mRNA translating multiple peptides including CAPA-DH and two CAPA-PVK (periviscerokinin) peptides to act on different receptors. Comparing the genes of SWD PK, CAPA, PKr and CAPA-DHr to four corresponding genes of D. melanogaster, SWD CAPA and the receptor are more similar to D. melanogaster than PK and the receptor. These data suggest that the CAPA gene could be evolutionally more conserved to have a common biological role in insects. In addition, the effect of Kozak sequences was investigated by the expression of the GPCRs with or without Kozak sequences in Sf9 insect cells. The Kozak sequenced PK receptor was significantly less active than the original (= no Kozak sequenced) receptor. Our results provide a knowledge for potential biological function(s) of PK and CAPA-DH peptides in SWD, and possibly offer a novel control method for this pest insect in the future.

CRISPR-Cas9 Mediated Gene-Silencing of the Mutant Huntingtin Gene in an In Vitro Model of Huntington's Disease.

Huntington's disease (HD) is a fatal neurodegenerative genetic disease characterized by a loss of neurons in the striatum. It is caused by a mutation in the Huntingtin gene (HTT) that codes for the protein huntingtin (HTT). The mutant Huntingtin gene (mHTT) contains extra poly-glutamine (CAG) repeats from which the translated mutant huntingtin proteins (mHTT) undergo inappropriate post-translational modifications, conferring a toxic gain of function, in addition to its non-functional property. In order to curb the production of the mHTT, we have constructed two CRISPR (clustered regularly interspaced short palindromic repeat)-Cas9 (CRISPR associate protein) plasmids, among which one nicks the DNA at untranslated region upstream to the open reading frame (uORF), and the other nicks the DNA at exon1-intron boundary. The primary goal of this study was to apply this plasmid into mesenchymal stem cells (MSCs) extracted from the bone-marrow of YAC128 mice, which carries the transgene for HD. Our results suggest that the disruption of uORF through CRISPR-Cas9 influences the translation of mHTT negatively and, to a lesser extent, disrupts the exon1-intron boundary, which affects the translation of the mHTT. These findings also revealed the pattern of the nucleotide addition or deletion at the site of the DNA-nick in this model.

Comparative analysis of contextual bias around the translation initiation sites in plant genomes.

Nucleotide distribution around translation initiation site (TIS) is thought to play an important role in determining translation efficiency. Kozak in vertebrates and later Joshi et al. in plants identified context sequence having a key role in translation efficiency, but a great variation regarding this context sequence has been observed among different taxa. The present study aims to refine the context sequence around initiation codon in plants and addresses the sampling error problem by using complete genomes of 7 monocots and 7 dicots separately. Besides positions -3 and +4, significant conservation at -2 and +5 positions was also found and nucleotide bias at the latter two positions was shown to directly influence translation efficiency in the taxon studied. About 1.8% (monocots) and 2.4% (dicots) of the total sequences fit the context sequence from positions -3 to +5, which might be indicative of lower number of housekeeping genes in the transcriptome. A three base periodicity was observed in 5' UTR and CDS of monocots and only in CDS of dicots as confirmed against random occurrence and annotation errors. Deterministic enrichment of GCNAUGGC in monocots, AANAUGGC in dicots and GCNAUGGC in plants around TIS was also established (where AUG denotes the start codon), which can serve as an arbiter of putative TIS with efficient translation in plants.

Tuning gene expression with synthetic upstream open reading frames.

We engineered short ORFs and used them to control the expression level of recombinant proteins. These short ORFs, encoding a two-amino acid peptide, were placed upstream of an ORF encoding a protein of interest. Insertion of these upstream ORFs (uORFs) resulted in suppression of protein expression. By varying the base sequence preceding the uORF, we sought to vary the translation initiation rate of the uORF and subsequently control the degree of this suppression. Using this strategy, we generated a library of RNA sequence elements that can specify protein expression over a broad range of levels. By also using multiple uORFs in series and non-AUG start codons, we were able to generate particularly low expression levels, allowing us to achieve expression levels spanning three orders of magnitude. Modeling supported a mechanism where uORFs shunt the flow of ribosomes away from the downstream protein-coding ORF. With a lower translation initiation rate at the uORF, more ribosomes "leak" past the uORF; consequently, more ribosomes are able to reach and translate the downstream ORF. We report expression control by engineering uORFs and translation initiation to be robust, predictable, and reproducible across all cell types tested. We propose control of translation initiation as a primary method of choice for tuning expression in mammalian systems.

Quantitative analysis of mammalian translation initiation sites by FACS-seq.

An approach combining fluorescence-activated cell sorting and high-throughput DNA sequencing (FACS-seq) was employed to determine the efficiency of start codon recognition for all possible translation initiation sites (TIS) utilizing AUG start codons. Using FACS-seq, we measured translation from a genetic reporter library representing all 65,536 possible TIS sequences spanning the -6 to +5 positions. We found that the motif RYMRMVAUGGC enhanced start codon recognition and translation efficiency. However, dinucleotide interactions, which cannot be conveyed by a single motif, were also important for modeling TIS efficiency. Our dataset combined with modeling allowed us to predict genome-wide translation initiation efficiency for all mRNA transcripts. Additionally, we screened somatic TIS mutations associated with tumorigenesis to identify candidate driver mutations consistent with known tumor expression patterns. Finally, we implemented a quantitative leaky scanning model to predict alternative initiation sites that produce truncated protein isoforms and compared predictions with ribosome footprint profiling data. The comprehensive analysis of the TIS sequence space enables quantitative predictions of translation initiation based on genome sequence.

A mutation in the Kozak sequence of GATA4 hampers translation in a family with atrial septal defects.

Atrial septal defect (ASD) is the most common congenital heart defect clinically characterized by an opening in the atrial septum. Mutations in GATA4, TBX5, and NKX2-5 underlie this phenotype. Here, we report on the identification of a novel -6 G>C mutation in the highly conserved Kozak sequence in the 5'UTR of GATA4 in a small family presenting with two different forms of ASD. This is the first time a mutation in the Kozak sequence has been linked to heart disease. Functional assays demonstrate reduced GATA4 translation, though the GATA4 transcript levels remain normal. This leads to a reduction of GATA4 protein level, consequently diminishing the ability of GATA4 to transactivate target genes, as demonstrated by using the GATA4-driven Nppa (ANF) promoter. In conclusion, we identified a mutation in the GATA4 Kozak sequence that likely contributes to the pathogenesis of ASD. In general, it points to the importance of accurate protein level regulation during heart development and emphasizes the need to analyze the entire transcribed region when screening for mutations.

Expression of soluble recombinant lipoxygenase from Pleurotus sapidus in Pichia pastoris.

The first heterologous expression of an iron-containing lipoxygenase from a basidiomycete in Pichia pastoris is reported. Five different expression constructs of the lipoxygenase gene LOX1 from Pleurotus sapidus were cloned and successfully transferred into P. pastoris SMD1168, but only one pPIC9K vector construct was functionally expressed. In this construct the vector-provided α-factor signal sequence was replaced by insertion of a second Kozak sequence between the signal sequence and the LOX1 gene. His(+) transformants were screened for their level of resistance to geneticin (G418). Lox1 was expressed under different culture conditions and purified using the N-terminal His-tag. Relative enzyme activity increased significantly 48h after methanol induction and was highest with 2mll(-1) inducer. The recombinant enzyme showed an optimal lipoxygenase activity at pH 7 and 30-35°C and a vmax like the wild-type enzyme.

Exploiting Systematic Engineering of the Expression Cassette as a Powerful Tool to Enhance Heterologous Gene Expression in Trichoderma reesei.

Many endeavors in expressing a heterologous gene in microbial hosts rely on simply placing the gene of interest between a selected pair of promoters and terminator. However, although the expression efficiency could be improved by engineering the host cell, how modifying the expression cassette itself systematically would affect heterologous gene expression remains largely unknown. As the promoter and terminator bear plentiful cis-elements, herein using the Aspergillus niger mannanase with high application value in animal feeds and the eukaryotic filamentous fungus workhorse Trichoderma reesei as a model gene/host, systematic engineering of an expression cassette was investigated to decipher the effect of its mutagenesis on heterologous gene expression. Modifying the promoter, signal peptide, the eukaryotic-specific Kozak sequence, and the 3'-UTR could stepwise improve extracellular mannanase production from 17 U/mL to an ultimate 471 U/mL, representing a 27.7-fold increase in expression. The strategies can be generally applied in improving the production of heterologous proteins in eukaryotic microbial hosts.

Optimization of extended Kozak elements enhances recombinant proteins expression in CHO cells.

In eukaryotes, the localization of small ribosomal subunits to mRNA transcripts requires the translation of Kozak elements at the starting site. The sequence of Kozak elements affects the translation efficiency of protein synthesis. However, whether the upstream nucleotide of Kozak sequence affects the expression of recombinant proteins in Chinese hamster ovary (CHO) cells remains unclear. In order to find the optimal sequence to enhance recombinant proteins expression in CHO cells, -10 to +4 sequences around ATG in 100 CHO genes were compared, and the extended Kozak elements with different translation intensities were constructed. Using the classic Kozak element as control, the effects of optimized extended Kozak elements on the secreted alkaline phosphatase (SEAP) and human serum albumin (HSA) gene were studied. The results showed that the optimized extended Kozak sequence can enhance the stable expression level of recombinant proteins in CHO cells. Furthermore, it was found that the increased expression level of the recombinant protein was not related with higher transcription level. In summary, optimizing extended Kozak elements can enhance the expression of recombinant proteins in CHO cells, which contributes to the construction of an efficient expression system for CHO cells.