Restriction Enzymes at NEB: Over 30 years of Innovation

Note: This paper was written in 2008.

The early days

In 1975, New England Biolabs (NEB) became the first company to sell restriction enzymes, thereby providing researchers with a key set of tools that proved invaluable for the early development of the biotechnology industry. Many of those early restriction enzymes had been discovered in my laboratory at Cold Spring Harbor Laboratory (CSHL). Summer visitors would stop by with a tube of their favorite DNA in their pocket, just to see if we had an enzyme that would convert it into some useful fragments. At that time, there was no commercial source for reagents, and I was unsuccessful in persuading CSHL to consider being that source and using the proceeds to fund basic research. A short time later, I met Dr. Donald Comb, who had recently started a company, New England Biolabs, based on this very same principle. We struck up a lasting partnership, and I was able to serve as a consultant to NEB, providing strains for many enzymes, while performing quality control experiments in my lab at CSHL prior to their commercial availability at NEB. Soon NEB was selling most of the known specificities and had begun a vigorous program to clone the genes for these enzymes in order to improve their quality and efficiency of production.

In addition, NEB embarked on an ambitious research program to find out as much as possible about restriction enzymes and their properties. That program continues today and we have arrived at the point where we have cloned, sequenced and characterized the genes for all but a handful of the restriction enzymes we sell. Along the way we have established wide-ranging academic collaborations to obtain structures for many of these enzymes. We also regularly provide purified enzymes and clones to our academic colleagues interested in researching these fascinating enzymes.

The discovery and screening process

As NEB grew and became known as a premier supplier of restriction enzymes, it was natural to begin screening microorganisms for new ones. This began in earnest in 1978, and over the years more than 500 restriction enzymes have been discovered by NEB scientists. This includes not only in-house scientists, but also many collaborating investigators from overseas. We have supported a series of screening labs in China, Cameroon, Vietnam, Nicaragua, Uganda and Portugal, which has helped these labs become established within their own countries. For the most part screening was accomplished by the original methods. This involved growing locally isolated novel organisms, breaking them open and testing either crude extracts or column fractions for activities able to give specific fragmentation patterns on bacteriophage lambda DNA. By testing candidate enzymes on several different DNAs, it was often possible to show a new activity even before detailed characterization had taken place. Today, almost half of the enzymes sold by NEB were discovered here.

Cloning and engineering of restriction enzymes

As NEB expanded its offering of restriction enzymes it soon became apparent that the demand for some enzymes was greater than the supply. Also, some enzymes were notoriously difficult to purify from native sources due to the presence of contaminating activities, such as non-specific nucleases, that routinely co-purified with the restriction enzyme. Since the fledgling biotechnology industry was focused on producing useful proteins by recombinant DNA methodology, it became clear that NEB should do the same for restriction enzymes. If overexpression could be achieved, purification protocols would be simplified. For many enzymes, overexpression has enabled NEB to reduce prices 5- to 20-fold. In cases where the native organism contained more than one restriction enzyme, the cloned version would guarantee greater purity with no cross-contamination. Furthermore, having access to cloned and sequenced genes would offer opportunities to increase our basic knowledge about these enzymes and lead to the possibility of improving their properties.

The first restriction-modification (R-M) system cloned at NEB was PstI. Soon to follow were EcoRI, MspI, HindIII and TaqαI. As the number of researchers working in this area grew, BamHI, FokI and BglII were cloned, as well as the important and useful 8-base cutters SfiI, NotI, and PacI. Some systems were cloned using bacteriophage to select for recombinants that expressed the restriction phenotype. Others were cloned using variations of an enzymatic method to select for recombinants that expressed the modification phenotype (1,2). In many cases, the R (restriction) and M (methylase) genes could be transferred simultaneously into E. coli. In other cases, it was necessary to clone them sequentially, first M and then R (3). Occasionally, additional non-cognate M genes were needed to provide complete protection from self-restriction. Cloning M genes alone was often problematic until it was discovered that E. coli synthesizes enzymes that specifically destroy modified DNA (4). Numerous obstacles were encountered in the course of cloning and over-expressing restriction enzymes – stemming mainly from the fact that each is unique – and even today they can still present exasperating problems. However, as a result of our extensive research activities in this area we now have unparalleled knowledge of the organization and properties of these systems – knowledge that has been directly applied to our ability to manufacture the very highest quality enzymes available.

One of the research focuses at NEB has been to clone and sequence the genes encoding all of the restriction enzymes that we sell. So far, this has been accomplished for 222 of 231 restriction enzymes; to date, over 180 are already commercially available*. This work has provided valuable information that we use when screening newly sequenced genomes for the presence of Type II restriction enzyme genes. It is often possible to identify new examples of well known specificities on the basis of sequence similarity, but more importantly it has enabled us to identify potential genes that might encode new specificities. This is now a key part of our screening efforts to find new restriction enzymes.

Once a restriction enzyme gene had been cloned, it became natural to try to change its properties by engineering. Until recently, changing specificity proved troublesome, and only a few non-useful changes had been described in the literature (5,6,7). One set of useful mutants that has been produced is able to selectively nick one strand but not the other (8). Many of these nicking endonucleases are now commercially available, including Nb.BbvCI and Nt.BstNBI. These novel enzymes are generated both by incapacitating or omitting individual subunits of heterodimeric restriction enzymes (9,10) or by preventing dimerization of others (11). Most recently, scientists at NEB have had great success in engineering Type IIG enzymes, a specific subtype of restriction enzymes. It is now possible to generate a wide variety of new sequence specificities in a directed fashion by discrete amino acid substitution (12). Additionally, fusions between different types of restriction enzymes have introduced new properties into them such as cleaving on both sides of their recognition sequence (13), while mutagenesis has had some dramatic effects on the star activity of many restriction enzymes (14).

Infidelity among the restriction enzymes

While most restriction enzymes cleave their recognition sequences with great fidelity, some are notorious for their propensity to cleave at secondary sites that are closely related to their cognate sites. One of the earliest observations of this promiscuous behavior was seen for EcoRI (cognate recognition site: GAATTC), which was found to cleave at sites differing from this “proper” site by one base (15). It was found that under certain buffer conditions, such as low ionic strength, low Mg2+ or the presence of organic solvents, sites such as NAATTC, GNATTC or GANTTC could be cleaved, albeit at reduced efficiency. This unwanted cleavage became known as star activity, and has been the bane of researchers looking for faithful cleavage ever since.

In addition to EcoRI, a number of other enzymes have been shown to exhibit star activity, including BamHI, PvuII and EcoRV (16). The conditions leading to this additional cleavage are now well documented. A quick glance at most catalogs containing restriction enzymes show that a rather large number of enzymes exhibit this unwanted behavior. Fortunately, under ideal buffer conditions, which can change substantially from one enzyme to another, star activity can be eliminated or at least greatly reduced in many cases. Since it is often desirable to perform a restriction digestion immediately following another reaction (i.e. ligation or amplification) or to use two restriction enzymes simultaneously, buffer conditions for one of these reactions is often far from ideal. In these circumstances, the need to change buffers can be time-consuming, may lead to sample loss or the desired products are formed at low yield. Recent innovations at NEB are poised to help.

Several major advances in restriction enzyme technology have been made that can alleviate problematic star activity. Researchers at NEB and the Indian Institute of Science in Bangalore described a mutant of KpnI that shows reduced star activity (17). Now, mutants of a number of other restriction enzymes have been prepared at NEB that can effectively eliminate or greatly reduce star activity for many commonly used enzymes. These newly released, high fidelity (HF™) enzymes are sold as separate products at the same low price as their wild type counterparts.

The current offering of HF enzymes with reduced star activity is only the beginning of improved products in the restriction enzyme field that we will be offering. Twenty-plus years of research has been rewarding in terms of knowledge acquired about these systems. From that knowledge has sprung an understanding of how these reagents might be improved for maximum utility. One of our goals is to only be selling restriction enzymes that are prepared from clones, thus ensuring unparalleled purity. Also, we will be offering additional engineered variants with altered and much improved performance and convenience. Once again our dedication to research, which is an integral part of our business philosophy, has paid off in better products and a deeper understanding of the products we sell.

* 2013 update: Over 250 of the 267 restriction enzymes are recombinant.


  1. Szomolanyi, I., Kiss, A., and Venetianer, P. (1980) Gene, 10, 219–225.
  2. Wilson, G.G. (1988) Gene, 74, 281–289.
  3. Howard, K.A., et al. (1986) Nucleic Acids Res. 14, 7939–7951.
  4. Raleigh, E.A. and Wilson, G.G. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 9070–9074.
  5. Whitaker, R.D., Dorner, L.F., and Schildkraut, I. (1999) J. Mol. Biol. 285, 1525–1536.
  6. Rimseliene, R., et al. (2003) J. Mol. Biol. 327, 383–391.
  7. Samuelson, J.C., et al. (2006) Nucleic Acids Res. 34, 796–805.
  8. Heiter, D.F., Lunnen, K.D. and Wilson, G.G. (2005) J. Mol. Biol. 348, 631–640.
  9. Bellamy, S.R.W., et al. (2005) J. Mol. Biol. 348, 641–653.
  10. Xu, S.Y., et al. (2007) Nucleic Acids Res. 35, 4608–4618.
  11. Xu, Y., Lunnen, K.D. and Kong, H. (2001) Proc. Natl. Acad. Sci. USA, 98, 12990–12995.
  12. Morgan, R.D., unpublished results.
  13. Zhang, P., et al. (2007) Prot. Engineering, Design and Selection. 20, 497–504.
  14. Zhu, Z. and Xu, S.Y., unpublished results.
  15. Polisky, B., et al. (1975) Proc. Natl. Acad. Sci. USA, 72, 3310–3314.
  16. Robinson, C.R., and Sligar, S.G. (1995) Proc. Natl. Acad. Sci. USA, 92, 3444–3448.
  17. Zhu, Z., Nagaraja, V, Saravanan, M. International Patent Office (2007). WO 200727464 A.
  18. Vincze, T., Posfai, J. and Roberts, R.J. (2003) Nucleic Acids Res. 31, 3688–3691.
From NEB Expressions Winter 2008, vol 2.4 By Richard J. Roberts, Chief Scientific Officer, NEB and 1993 Nobel Laureate in Physiology or Medicine.