Expression Libraries

 An expression library is a collection of DNA molecules that can be expressed en masse through enzymatic methods to yield a population of DNA, RNA, or protein products that can be screened for individuals displaying desirable properties. The DNA sequences within the library can be of natural or synthetic origin. The advantage of screening expression libraries constructed from the genome or complementary DNA  is that the expression products derived are usually biologically relevant. In contrast, synthetic DNA expression libraries may yield expression products that are not found in nature. However, there are two major advantages to screening synthetic DNA expression libraries. First, the expression products that are selected for a particular purpose may perform better than natural biopolymers. Second, a great deal of information is obtained by screening synthetic libraries of known composition, because the distribution of sequences selected compared with those not selected provides structure–activity relationship (SAR( data describing molecular interactions characteristic of the target being screened.

Expression libraries contain genomic or cDNA sequences cloned into expression vectors, which are plasmids or bacteriophage genomes that have been engineered to contain regulatory elements governing expression of foreign DNA inserts. For example, vectors containing a promoter sequence of T7 phage will initiate RNA transcription at a fixed position and direction. Once transcription is initiated, the T7 RNA polymerase will continue making an RNA copy of the DNA coding strand, including any foreign DNA that is encountered, in progressive fashion until a T7 terminator sequence is read (by convention, the coding strand is the top strand in a 5′ to 3′, left to right, double-stranded gene depiction). The resulting polymer that is formed is thus an RNA transcript of the DNA template. Such a DNA library can only be expressed in cells that also contain the enzyme T7 RNA polymerase. Typically, the gene encoding the T7 RNA polymerase is engineered to be inducible by the addition of a soluble sugar analogue, such as IPTG, so that library RNA expression can be co-induced at appropriate times in the screening process. RNA expressed from DNA libraries may be the intended chemical entity to be screened, or the RNA molecules may serve as templates for the translation of protein products. If the expressed RNA molecules contain suitable prokaryotic and/or eukaryotic ribosome recognition sequences, ribosome-mediated protein biosynthesis will initiate at an AUG start codon. The result will be a protein molecule encoded by the original DNA template. In general, it is easiest to screen protein expression libraries when the transcription and translation occur in vivo using a suitable host cell. However, in certain circumstances it may be advantageous to perform the protein expression in vitro, using cell-free translation extracts that contain the requisite protein synthesis machinery. In either case, the population of expressed proteins can be screened for any of a range of phenotypic properties, including antibody reactivity, affinity to a target molecule, or catalysis.

 The ability to engineer, at the DNA level, additional useful sequences into protein reading frames provides a powerful means of generating fusion proteins with desirable properties. Modern expression libraries contain many types of tags or fusion domains that facilitate library screening and downstream manipulation of individual clones. There are at least four types of protein fusion moieties commonly available. This list is not intended to be exhaustive but instead to illustrate the power of the technology. First, epitope tags are typically short amino acid segments fused to either the N- or C-terminus of the cloned protein. An epitope tag provides ready-made immunoreactivity to the cloned gene product through commercially available antibodies. Epitope tags are useful in both detection and purification of cloned gene products. Second, like epitope tags, affinity tags are amino acid sequences fused to either terminus of the cloned protein. The most common use of affinity tags is in affinity purification, although detection can also be performed through the tag. One commonly used affinity tag is the 6-His tag, which is simply a sequence of six histidine residues that bind with high affinity and specificity to nickel; therefore, 6-His fusion proteins can be purified from all other cellular proteins by affinity chromatography over nickel-agarose (1). Third, domain tags are independent units of structure and function that maintain or enhance certain activities of the proteins from which they were derived. Domain tags are useful for protein purification and detection. Common examples include glutathione-S-transferase (GST), which enables facile purification over glutathione-agarose resins (2), and green fluorescent protein (GFP), which is an autofluorescent protein that can be monitored in both quantitative and localization assays (3). Fourth, cleavage sequences are commonly introduced between the fusion tags and the cloned genes so that the tags can be removed following affinity purification. Cleavage elements are most commonly proteinase recognition sequences, although self-cleaving protein domains called inteins are becoming increasingly available (4).

A special type of cDNA expression library called the two-hybrid system (5) deserves special mention because it illustrates a powerful coupling of molecular engineering with combinatorial approaches to investigating protein function. The purpose of screening two hybrid libraries is to discover proteins that interact with a protein of interest. The system is termed “two-hybrid” because two separate plasmids have been engineered to express hybrid versions of transcriptional activators such as the yeast GAL4 protein. Normally, the GAL4 protein contains both a DNA-binding domain and a transcriptional activation domain. When intact GAL4 binds to its DNA recognition sequence, the transcriptional activation domain promotes transcription of adjacent gene sequences. In the two-hybrid system, the GAL4 DNA binding domain is expressed as a hybrid with the protein of interest, termed the “bait”. A second vector containing the GAL4 transcriptional activation domain is used to construct a library of cDNAs such that each cDNA is expressed as a hybrid protein fused to the GAL4 transcriptional activation domain. Neither of the hybrid GAL4 domains alone can promote transcription of a reporter gene, nor do they interact spontaneously. However, if one of the proteins expressed as a fusion with the GAL4 transcriptional activation domain happens to form a protein-protein interaction with the bait protein, the two GAL4 domains are brought into proximity where they can activate expression of a reporter gene. Many variations of this powerful method have since appeared, including two-hybrid systems for mammalian cells (6) and three-hybrid systems for detecting protein-nucleic acid interactions (7).

 The other major type of expression library uses synthetic DNA to express DNA, RNA, or protein products for screening. Expression of RNA and protein can be mediated through cloning vectors in similar fashion to the cDNA expression libraries. Examples of cloned synthetic DNA expression libraries include phage display and protein engineering projects in which portions of a protein-coding sequence are randomized synthetically to derive novel proteins with altered properties. Alternatively, the synthetic DNA library may be maintained as oligonucleotides that are replicated through PCR or some other enzymatic amplification strategy. Expression of these types of libraries always occurs in vitro and usually employs a synthetic phage T7 promoter to mediate expression of RNA. Thus, RNA copies of the DNA library are produced upon incubation with purified T7 RNA polymerase and ribonucleoside triphosphate (rNTP) building blocks. The resulting single-stranded RNA can be screened for affinity to specific molecular targets (RNA aptamer libraries) or for catalytic activity (ribozyme libraries). Similarly, single-stranded DNA copies can be prepared from DNA aptamer libraries by removing one of the two DNA strands, which has been previously tagged with biotin, by denaturation and affinity chromatography over a streptavidin matrix. DNA aptamers are similar in many ways to RNA aptamers, but they are more resistant to chemical and enzymatic degradation. Finally, peptide aptamers can in principle be prepared by in vitro translation of RNA transcripts from oligonucleotide libraries, although this method is not commonly employed.

References

1. J. Schmitt, H. Hess, and H. G. Stunnenberg (1993) Mol. Biol. Rep. 18, 223–230. 

2. D. B. Smith and K. S. Johnson (1988) Gene 67, 31–40. 

3. A. Crameri, E. A. Whitehorn, E. Tate, and W. P. C. Stemmer (1996) Nat. Biotechnol. 14, 315–319.

4. S. Pietrokovski (1994) Protein Sci. 3, 2340–2350. 

5. C. T. Chien, P. L. Bartel, R. Sternglanz, and S. Fields (1991) Proc. Natl. Acad. Sci. USA 88, 9578-9582.

6.Y. Luo, A. Batalao, H. Zhou, and L. Zhu (1997) Biotechniques 22, 350–352. 

7. D. J. SenGupta, B. Zhang, B. Kraemer, P. Pochart, S. Fields, and M. Wickens (1996) Proc. Natl. Acad. Sci. USA 93, 8496–8501.