Introduction to Expression by Fusion UNIT 16.

4A
Protein Vectors
Expressionthe directed synthesis of a for- (UNITS 16.5-16.8) describe the use of vectors that
eign geneis often the logical next step for express -galactosidase and trpE fusions, mal-
researchers who have isolated a gene and want tose-binding protein (MBP) fusions, glu-
to study the protein it encodes. During the early tathione-S-transferase (GST) fusions, and thio-
days of recombinant DNA technology, it was redoxin (Trx) fusions. These carrier regions
thought that a strong promoter and a start codon often can be exploited in purifying the protein,
at the beginning of the gene would be sufficient either with antibodies or with an affinity puri-
for good expression in Escherichia coli. Since fication specific for that carrier protein. Alter-
then it has been learned that the requirements natively, unique physical properties of the car-
for efficient translation are a good deal more rier protein (e.g., heat stability) can be exploited
complicated. In addition to a promoter and a to allow selective purification of the fusion
start codon, good expression requires that the protein. In addition, some carrier proteins such
mRNA encoding the protein to be expressed as MBP and Trx can be selectively released
contain a ribosome-binding site that is not from intact cells by osmotic shock or
blocked by mRNA secondary structure. The freeze/thaw procedures, even though they re-
level of expression is also affected by codon side in different cellular compartments. Often,
preferences, especially in the second codon of proteins fused to these carriers can be separated
the gene (Stormo et al., 1982), and may be from the bulk of intracellular contaminants by
affected by the coding sequence in other ways taking advantage of this attribute.
that are not yet well understood (UNIT 16.1). In There are three problems often encountered
virtually all cases, these problems can be solved when expressing fusion proteins: solubility of
by altering the sequence preceding the start the expressed protein, stability of the expressed
codon, and/or by making changes in the 5 end protein, and presence of the carrier protein. The
of the coding sequence that do not change the first two problems are often encountered with
protein sequence, taking advantage of the de- both fusion and nonfusion expression systems
generacy of the genetic code. (UNIT 16.1), while the third is unique to fusion
However, it is often quicker to solve these systems.
problems by making fusions between genes. In
this approach the cloned gene is introduced into SOLUBILITY OF THE EXPRESSED
an expression vector 3 to a sequence (carrier PROTEIN
sequence) coding for the amino terminus of a The high-level expression of many proteins
highly expressed protein (carrier protein). The can lead to the formation of inclusion bodies,
carrier sequence is often from an E. coli gene, very dense aggregates of insoluble protein and
but it can be from any gene that is strongly RNA that contain most of the expressed protein
expressed in E. coli. The carrier sequence pro- (Schein, 1989). Precipitation of a protein into
vides the necessary signals for good expres- inclusion bodies sometimes can work to ones
sion, and the expressed fusion protein contains advantage, because inclusion bodies are insol-
an N-terminal region encoded by the carrier. In uble and dense, and can be purified relatively
such vectors, the portion of the fusion protein easily by centrifugation (UNIT 16.5). In addition,
encoded by the carrier can be as small as one some proteins that are degraded when ex-
amino acid (UNIT 16.3; Amann and Brosius, pressed in the soluble fraction are quite stable
1985), although expression from such vectors as inclusion bodies. Once purified, protein in
can still be subject to problems caused by the inclusion bodies can be solubilized by denatu-
coding sequence of the expressed protein. Per- ration with guanidineHCl or urea, and then
haps more typical examples of short carrier can often be refolded by dialyzing away the
sequences are those contained in the trpE vec- denaturant. A problem, however, with denatu-
tors (UNIT 16.5) or the cII vectors (Nagai and ration/renaturation is that the yield of properly
Thgersen, 1987). refolded protein is variable and sometimes
The carrier sequence can also code for an quite low; some proteins, especially large
entire functional moiety or even for an entire ones, cannot be properly refolded at all (see
protein. For example, the following four units UNIT 16.5).
Protein
Expression
Contributed by Paul Riggs and Edward R. La Vallie 16.4.1
Current Protocols in Molecular Biology (1994) 16.4.1-16.4.4
Copyright 2000 by John Wiley & Sons, Inc. Supplement 28
 If expression of a particular fusion protein dation of unstable proteins (Baker et al., 1984;
produces insoluble aggregates and a soluble I. Hall and P. Riggs, unpub. observ.). Similarly,
protein is required, there are several things to the degP mutant has been shown to stabilize
try. One important variable is temperature; for fusion proteins in the periplasm (Strauch and
reasons not well understood, higher tempera- Beckwith, 1988) and ompT mutants have
tures (37 and 42C) promote inclusion-body proven useful in preventing cleavage between
formation and lower temperatures (30C) in- exposed basic residues (e.g., Arg-Arg) in sev-
hibit it (Bishai et al., 1987; Schein, 1989). eral nonfusion proteins during preparation of
Another variable is the level of expression; crude extracts (Grodberg and Dunn, 1988;
sometimes lowering the expression level can Sugimura and Higashi, 1988). Finally, the sta-
increase the proportion of protein that is sol- bility of a particular fusion can vary even
uble. A third variable is the strain background among different wild-type lab strains, per-
of the cells bearing the expression vector; large haps due to uncharacterized differences in pro-
differences in the proportion of a particular tease levels among the strains (I. Hall, P. Riggs,
expressed protein that is soluble are seen M. Southworth, S. Levitt, and F. Perler, unpub.
among different strains (M. Southworth, S. observ.).
Levitt, and F. Perler, unpub. observ.; it is not
known which of the genetic differences be- CLEAVAGE OF FUSION PROTEINS
tween the strains is responsible for the differ- TO REMOVE THE CARRIER
ences in solubility). Finally, it is worth noting The use of fusion proteins is growing rapidly
that changes in the carrier protein can affect for the many reasons described above. The
the solubility of an expressed fusion protein various systems described in the following
(La Vallie et al., 1993). units have been used to produce many different
kinds of proteins ranging from enzymes and
STABILITY OF THE EXPRESSED growth factors to transmembrane receptors and
PROTEIN DNA binding proteins. Often it is advantageous
Stability problems are often encountered to remove the carrier protein moiety from the
when foreign proteins, especially eukaryotic protein of interest to facilitate biochemical and
proteins, are expressed in E. coli. The carrier functional analyses. Several methods for site-
protein can sometimes stabilize an expression specific cleavage of fusion proteins have been
fusion protein (Lee et al., 1984). Sometimes, developed (UNIT 16.4B). The choice of method is
however, the expressed protein is degraded but usually determined by the composition, se-
the carrier protein is not. Moreover, fusion quence, and physical characteristics of the par-
proteins are sometimes cleaved in vivo at the ticular protein. Chemical cleavage of fusion
fusion joint between the carrier and expressed proteins can be accomplished with reagents
portions of the fusion, which obviously creates such as cyanogen bromide (Met, Itakura et
problems if the carrier protein is to be used as al., 1977), 2-(2-nitrophenylsulphenyl)-3-
an aid in purification. These facts about fusion methyl-3-bromoindolenine (BNPS-skatole,
proteins are consistent with a model in which Trp, Dykes et al., 1988), hydroxylamine (As-
the carrier and the rest of the protein form nGly, Bornstein and Balian, 1977), or low
independent domains. In this view, it can be pH (AspPro, Szoka et al., 1986). Chemical
imagined that there are cases where the carrier cleavage procedures tend to be inexpensive and
domain folds correctly and the expressed pro- efficient, and often can be accomplished under
tein does not (and is degraded). There are also denaturing conditions to cleave otherwise in-
cases where both domains fold correctly but the soluble fusion proteins (Szoka et al., 1986).
joint region between them is sensitive to one or However, their use is hampered by the likely
more E. coli proteases. occurrence of cleavage sites in the protein of
Approaches that have been used to stabilize interest, along with the propensity for side
fusion proteins are generally the same as those reactions that result in unwanted modifications
used to stabilize nonfusion proteins. One to the protein. As an alternative to chemical
method is to arrange for the fusion protein to methods, enzymatic cleavage procedures are
be expressed as insoluble aggregates. Another desirable for their relatively mild reaction con-
method is to use E. coli strains deficient in ditions and, most importantly, for the high de-
known proteases. For example, a lon htpR dou- gree of specificity exhibited by some proteases
Introduction to ble-mutant strainwhich is deficient in several commonly used for this purpose. Among the
Expression by
Fusion Protein cytoplasmic proteasesshows reduced degra- useful enzymes are factor Xa (Nagai and
Vectors

16.4.2
Supplement 28 Current Protocols in Molecular Biology
Thgersen, 1984, 1987; Gardella et al., 1990), Gardella, T.J., Rubin, D., Abou-Samra, A.-B., Keut-
thrombin (Smith and Johnson, 1988; Gearing mann, H.T., Potts, J.T. Jr., Kronenberg, H.M.,
and Nussbaum, S.R. 1990. Expression of human
et al., 1989), enterokinase (Dykes et al., 1988;
parathyroid hormone-(1-84) in Escherichia coli
LaVallie et al., 1993), renin (Haffey at al., as a factor X cleavable fusion protein. J. Biol.
1987), and collagenase (Germino and Bastia, Chem. 265:15854-15859.
1984). All of these enzymes have extended Gearing. D.P., Nicola, N.A., Metcalf, D., Foote, S.,
substrate recognition sequences (up to 7 amino Willson, T.A., Gough, N.M., and Williams, R.L.
acids in the case of renin), which greatly re- 1989. Production of leukemia factor in Es-
duces the likelihood of unwanted cleavages cherichia coli by a novel procedure and its use
in maintaining embryonic stem cells in culture.
elsewhere in the protein. Of the above-men-
Bio/Technology 7:1157-1161.
tioned proteases, factor Xa and enterokinase are
Germino, J. and Bastia, D. 1984. Rapid purification
most useful in this application because they
of a gene product by genetic fusion and site-spe-
cleave on the carboxy-terminal side of their cific proteolysis. Proc. Natl. Acad. Sci. U.S.A.
respective recognition sequences, allowing the 81:4692-4696.
release of fusion partners containing their Grodberg, J. and Dunn, J.J. 1988. ompT encodes the
authentic amino-termini. Escherichia coli outer membrane protease that
UNITS 16.5, 16.6, 16.7 & 16.8 describe five different cleaves T7 RNA polymerase during purification.
fusion protein vector systems; of these, only J. Bacteriol. 170:1245-1253.
three include recognition sites for interdomain Haffey, M.L., Lehman, D., and Boger, J. 1987. Site-
cleavage. The MBP fusion system (UNIT 16.6) specific cleavage of a fusion protein by renin.
DNA 6:565-571.
provides a factor Xa cleavage site. The GST
fusion system (UNIT 16.7) includes vectors that Itakura, K., Hirose, T., Crea, R.M, Riggs, A.D.,
Heyneker, H.L., Bolivar, F., and Boyer, H.W.
contain either a thrombin cleavage site, a factor
1977. Expression in E. coli of a chemically syn-
Xa cleavage site, or an Asp-Pro acid cleavage thesized gene for the hormone somatostatin. Sci-
site. The Trx fusion system (UNIT 16.8) uses an ence 198:1053-1056.
enterokinase cleavage site. UNIT 16.4B describes LaVallie, E.R., Rehemtulla, A., Racie, L.A.,
fusion protein cleavages in detail, including DiBlasio, E.A., Ferenz, C., Grant, K.L., Light,
specific protocols for cleaving fusion proteins A., and McCoy, J.M. 1993. Cloning and func-
produced with each of the aforementioned vec- tional expression of a cDNA encoding the cata-
lytic subunit of bovine enterokinase. J. Biol.
tor systems, along with methodologies for the Chem. 268:23311-23317.
site-specific cleavage of proteins using various
Lee, N., Cozzikorto, J., Wainwright, N., and Testa,
chemical reagents.
D. 1984. Cloning with tandem gene systems for
high level gene expression. Nucl. Acids Res.
Literature Cited 12:6797-6812.
Amann, E. and Brosius, J. 1985. ATG vectors for Nagai, K. and Thgersen, H. C. 1984. Generation of
regulated high-level expression of cloned genes -globin by sequence-specific proteolysis of a
in Escherichia coli. Gene 40:183-190. hybrid protein produced in Escherichia coli. Na-
Baker, T.A., Grossman, D., and Gross, C.A. 1984. ture 309:810-812.
A gene regulating the heat shock response in Nagai, K. and Thgersen, H. C. 1987. Synthesis and
Escherichia coli also affects proteolysis. Proc. sequence-specific proteolysis of hybrid proteins
Natl. Acad. Sci. U.S.A. 81:6779-6783. produced in Escherichia coli. Methods Enzymol.
Bishia, W.R., Rappuoli, R., and Murphy, J.R. 1987. 153:461-481.
High-level expression of a proteolytically sensi- Schein, C.H. 1989. Production of soluble recombi-
tive diphtheria toxin fragment in Escherichia nant proteins in bacteria. Bio/Technology
coli. J. Bacteriol. 169:5140-5151. 7:1141-1149.
Bornstein, P. and Balian, G. 1977. Cleavage at Asn- Smith, D.B. and Johnson, K.S. 1988. Single-step
Gly bonds with hydroxylamine. Methods Enzy- purification of polypeptides expressed in Es-
mol. 47:132-145. cherichia coli as fusions with glutathione S-
Dykes, C.W., Bookless, A.B., Coomber, B.A., No- transferase. Gene 67:31-40.
ble, S.A., Humber, D.C., and Hobden, A.N. Stormo, G.D., Schneider, T.D., and Gold, L. 1982.
1988. Expression of atrial natriuretic factor as a Characterization of translation initiation sites in
cleavable fusion protein with chloramphenicol E. coli. Nucl. Acids Res. 10:2971-2996.
acetyltransferase in Escherichia coli. Eur. J. Bio-
chem. 174:411-416.

Protein
Expression

16.4.3
Current Protocols in Molecular Biology Supplement 28
 Strauch, K.L. and Beckwith, J. 1988. An Es-
cherichia coli mutation preventing degradation Contributed by Paul Riggs
of abnormal periplasmic proteins. Proc. Natl. New England Biolabs
Acad. Sci. U.S.A. 85:1576-1580. Beverly, Massachusetts
Sugimura, K. and Higashi, N. 1988. A novel outer-
membrane-associated protease in Escherichia Edward R. La Vallie
coli. J. Bacteriol. 170:3650-3654. and John M. McCoy
Szoka, P.R., Schreiber, A.B., Chan, H., and Murthy, Genetics Institute
J. 1986. A general method for retrieving the Cambridge, Massachusetts
components of a genetically engineered fusion
protein. DNA 5:11-20.

Introduction to
Expression by
Fusion Protein
Vectors

16.4.4
Supplement 28 Current Protocols in Molecular Biology