1

BIOSENSOR TECHNOLOGY APPLIED TO

HYBRIDIZATION ANALYSIS AND MUTATION DETECTION

PETER NILSSON

ROYAL INSTITUTE OF TECHNOLOGY

DEPARTMENT OF BIOTECHNOLOGY

STOCKHOLM 1998

4

Peter Nilsson (1998) Doctoral DissertationBiosensor technology applied to hybridization analysis and mutation detection.Dept. of Biotechnology, KTH - Royal Institute of Technology, 100 44 Stockholm, Sweden.ISBN 91-7170-322-5

ABSTRACT

This thesis demonstrates the application of biosensor technology for molecular biology

investigations, utilizing a surface plasmon resonance based optical device for mass sensitive

detection of biomolecular interactions at a chip surface. Oligonucleotide model systems were

designed for analysis of the action of DNA manipulating enzymes. DNA ligation, DNA

cleavage and DNA synthesis could be quantitatively monitored in real-time. A protocol for

DNA minisequencing was also established, based on prevention of chain elongation by

incorporation of chain-terminators. Determinations of affinities for short oligonucleotides

hybridizing to an immobilized target, were performed with various sequence content, length,

temperature and degree of complementarity. The decrease in affinity for hybridizations

involving mismatch situations was found to be strongly dependent on the relative position of

the mismatch. Interestingly, also end-mismatches were clearly detectable. The stabilization

effect achieved upon co-hybridization of two adjacently annealing short oligonucleotide

modules (modular primer effect) was also investigated for different module combinations and

hybridization situations. The modular concept of hybridizations was subsequently

demonstrated to result in enhanced capture of single stranded PCR products. The sequence-

based DNA analysis, first introduced with oligonucleotide model systems, was extended to the

scanning and screening for mutations in PCR amplified DNA from clinically relevant samples.

Several different formats were investigated, either with the PCR products immobilized on the

chip and oligonucleotides injected or vice versa. Again, mismatch discrimination could be

observed for wild type and mutant specific oligonucleotides hybridizing to the targets. The

experimental set-up for mutation detection was further developed by the introduction of a

subtractive mismatch sensitive hybridization outside the instrument and a subsequent

determination of the relative amounts of remaining oligonucleotides with analytical biosensor

monitoring of hybridizations between fully complementary oligonucleotides. In conclusion, the

applied technology was found to be a suitable tool for a wide range of molecular biology

applications, with emphasis on hybridization analysis and mutation detection.

Keywords: biosensor, genosensor, surface plasmon resonance, hybridization, nucleic acids,

oligonucleotide, mutation detection, mismatch discrimination, modular

5

LIST OF MAIN REFERENCES

This thesis is based on the following reports, which will be referred to in the text by their

Roman numerals:

I Nilsson, P., Persson, B., Uhlén, M. and Nygren, P-Å. (1995).

Real-time monitoring of DNA manipulations using biosensor technology.

Analytical Biochemistry 224: 400-408. A

II Persson, B., Stenhag, K., Nilsson, P., Larsson, A., Uhlén, M. and Nygren, P-Å. (1997).

Analysis of oligonucleotide probe affinities using surface plasmon resonance: a means for

mutational scanning.

Analytical Biochemistry 246: 34-44.A

III Nilsson, P., O’ Meara, D., Edebratt, F., Persson, B., Uhlén, M., Lundeberg, J. and

Nygren, P-Å. (1998)

Quantitative investigation of the modular primer effect for DNA and PNA hexamers.

(submitted)

IV O’ Meara, D., Nilsson, P., Nygren, P-Å., Uhlén, M. and Lundeberg, J. (1998).

Capture of single-stranded DNA assisted by oligonucleotide modules.

Analytical Biochemistry 255:195-203. A

V Nilsson, P., Persson, B., Larsson, A., Uhlén, M. and Nygren, P-Å. (1997).

Detection of mutations in PCR products from clinical samples by surface plasmon resonance.

Journal of Molecular Recognition 10:7-17. B

VI Nilsson, P., Larsson, A., Lundeberg, J., Uhlén, M. and Nygren, P-Å. (1998).

Mutational scanning of PCR products by subtractive oligonucleotide hybridization analysis.

BioTechniques (in press) C

The articles are reproduced with the permission of the copyright holders;©Academic Press, Inc. A, ©John Wiley & Sons, Ltd. B, ©Eaton Publishing Inc. C

6

TABLE OF CONTENTS

INTRODUCTION

SEQUENCE-BASED DNA ANALYSIS 8

HUMAN GENOME PROJECT 10

MUTATION DETECTION 11

Scanning for unknown mutations 12

Screening for known mutations 13

DNA arrays 15

Sequencing by hybridization 15

DNA chips 17

BIOSENSORS 19

Electrochemical transducers 19

Amperometric 19

Potentiometric 20

Conductimetric 21

Mechanical transducers 22

Piezoelectric 22

Thermal transducers 23

Calorimetric 23

Optical transducers 24

Evanescent wave 24

Surface plasmon resonance 24

Optical waveguiding 28

Other optical devices 29

7

TABLE OF CONTENTS

RESULTS AND DISCUSSION

INTRODUCTION 31

REAL TIME MONITORING OF DNA MANIPULATING ENZYMES (I)

32

DNA assembly using DNA ligase 32

Endonuclease cleavage 33

DNA synthesis by DNA polymerases 33

MISMATCH DISCRIMINATION IN OLIGONUCLEOTIDES (I, II) 34

Minisequencing (I) 34

Hybridization analysis by determination of equilibrium affinities (II) 35

Mutational scanning and influence of mismatch position (II)

36

ANALYSIS OF MODULAR HYBRIDIZATIONS (III, IV) 38

Increased stability by modular hybridizations (III) 38

The use of oligonucleotide modules for capture (IV) 40

MUTATION DETECTION IN CLINICAL SAMPLES (V, VI) 41

Mutation detection on immobilized or injected PCR amplicons (V) 41

Mutation detection by subtractive hybridizational scanning (VI) 43

CONCLUDING REMARKS 45

ABBREVIATIONS 46

ACKNOWLEDGEMENTS 47

REFERENCES 48

APPENDIX

Main references I - VI

8

INTRODUCTION

SEQUENCE-BASED DNA ANALYSIS

equence-based DNA analysis is today a corner-stone in many clinical and research

fields, such as biology, biotechnology, evolutionary studies, forensic science and a

wide variety of branches within the field of medicine. The analysis can roughly be

divided into two different areas; sequence specific detection or sequence determination. The

first type of analysis is aiming at detecting the presence of a certain nucleotide sequence in a

complex background, either for the purpose of identification, localization or just presence.

Sequence determination is more advanced and is aiming at identifying a specific nucleotide at a

specific position or more commonly, determine the whole sequence, i.e. the identities and

order of nucleotides in a DNA fragment.

The constantly increasing importance of sequence-based DNA analysis can not be argued by

anyone today. There has been an enormous development of technology for this purpose, since

the landmark paper by Avery et al in 1944, where it for the first time was suggested that DNA

and not protein is the hereditary material. Another more than obvious milestone is the Watson

and Crick paper in 1953, describing the double stranded helical nature of DNA and the

principle of the basepairing.

There have of course been many scientific breakthroughs related to DNA analysis during the

last decades, but three out of the major technical and methodological achievements are here

selected as very important for the present sequence-based DNA analysis and many of the more

recent developed techniques have utilized one or more of those three as a methodological

starting point.

E. M. Southern described in 1975 how specific nucleic acids could be identified in complex

mixtures using a technique later termed Southern blotting. In its typical format, the DNA to be

analyzed is fragmented with e.g. a restriction enzyme. The fragments are separated according

to size by gel electrophoresis and subsequently transferred to a filter or membrane. A labeled

DNA probe is hybridized, allowing the detection of its complementary sequence present on the

filter. This sequence specific hybridizational detection of nucleic acids and variants thereof, is

still a major component in many of the later developed techniques for the analysis of nucleic

acids.

S

9

A Nobel Prize awarded technique for determining the sequence of a DNA fragment

enzymatically was presented by Sanger et al in 1977. The method is now called Sanger or

dideoxy sequencing. A labeled primer is hybridized to a single stranded version of the DNA to

be sequenced. Added DNA polymerase extends the primer in the presence of the four

nucleotides and one chain-terminating dideoxynucleotide, which also can be an alternative

molecule to label. The terminator will be incorporated in a fraction of the fragments when the

nucleotide in the template is complementary to the dideoxynucleotide and prevent any further

extension of that fragment. After completed DNA synthesis, the reaction mixes from the use of

different terminators will contain fragments of different lengths, where each length corresponds

to a nucleotide position in the template. The fragments in the four reactions (one for each

dideoxynucleotide) are separated by gel electrophoresis and the sequence can be determined by

the order of the fragments from the four reaction mixes.

The third highlighted epoch-making, in the context of biotechnology, technique is the

polymerase chain reaction (PCR) (Mullis, 1986; 1987; Saiki, 1985; 1988), which also was

awarded the 1993 Nobel Prize in chemistry (K. B. Mullis). With the PCR came the ability to

selectively and exponentially amplify a desired DNA sequence present in complex mixtures and

produce millions of copies of it. "It is like finding a needle in a haystack and then make a new

stack of that needle" is a metaphor frequently used to describe PCR. The principle is simple

and in its simplest form based on enzymatic DNA synthesis in three-step temperature cycles

involving, denaturation of double strands, annealing of primers and finally, polymerase-

dependent primer extension. The procedure is repeated for approximately 30 cycles, leading to

a theoretical accumulation of approximately 230 copies of the desired fragment. The inventive

step was to combine a thermostable enzyme with the thermocycling and thereby avoiding the

need to add new enzyme after each cycle, which also led to increased primer specificity by

enabling a higher annealing temperature.

Both the PCR technology and the Sanger sequencing method have, in many aspects,

continuously been further developed and also several different variants of those techniques

have arisen. The triumphant introduction of the Sanger sequencing and the PCR amplification

and also the power of the combination of these techniques together with progress in

automation have led to a dramatic increase in sequencing capacity. Large scale sequencing

projects could be accomplished and with the increasing speed at which DNA sequence

information was generated, it was realized that the whole human genome would be possible to

sequence and an international human genome project was initiated.

10

HUMAN GENOME PROJECT

The human genome project (HGP) is coordinated and mainly funded by the US Department of

Energy (DOE) and the US National Institute of Health (NIH) and was formally started in 1990

after five years of debate and planning. Also several Japanese and European consortiums, such

as the Welcome Trust, UK, is involved in funding HGP. In 1988 the US National Academy of

Science reported that the complete human genome could be mapped and sequenced within 15

years. According to the original plans, the project would then be finished by the year 2005. A

5-year report in 1995 although stated that the project was ahead of schedule and that the goal

of having all human genes (approximately 80 000) discovered and the sequence of the

approximately 3 billion DNA bases determined, could be reached sooner than first estimated.

Several other genomes than the human are also included in HGP, as a way to increase the

knowledge of genome organization and for comparative genomic analysis. Among the non-

human organisms are several of the most studied and frequently used for experimental

manipulations, such as Saccharomyces cerevisiae (yeast), Escherichia coli (bacteria),

Drosophila melanogaster (fruitfly), Caenorhabditis elegans (nematode), Arabidopsis thaliana

(plant) and Mus musculus (mouse). In September 1998, 16 complete microbial genomes had

been completely sequenced, including S. cerevisiae and E. coli, and the complete sequencing

of the 97 Mb genome of C. elegans is predicted to be finished before the end of 1998. The

human genome had at the same time been sequenced to roughly 7 % (207 Mb / 3 Gb), if only

contigs longer than 1 kb are counted (116 Mb with 100 kb cut-off). The rate of sequencing has

been doubled the last two years and the estimation for the end of 1998 is 320 Mb of finished

sequence (Venter, 1998), corresponding to almost 11% of the genome.

HGP has greatly contributed to the increased rate of production of high-resolution maps,

which have been very important for the accelerated pace of discovery and identification of

disease related genes. The mapping and sequencing has proceeded in parallel. Sequence-based

gene discovery is although more and more replacing the map-based discovery by the shift from

structural to functional genomics with the aid of comparative genomics (McKusick, 1997).

The impact from this enormous sequencing and mapping effort on biology and medicine

research will probably lead to a paradigm shift, where some information of all genes are

available in databases. The databases themselves will be an important research tool for many

scientists and a lot of knowledge of genomic function can be gained directly from the

databases. Identification of sequence homology between human and other species, such as

11

yeast, can help define the function of genes. The increasingly high rate of genes being

identified, sequenced and given a function, will for certain be important for the analysis of

genes related to the development of diseases. More than 600 allelic variations of genes with

known disease-producing mutations have been catalogued (McKusick, 1997).

Figure 1. Human genomic sequences in the EMBL database. The columns represent the accumulated amountof sequence with contigs of at least 1 kb. The line represents the percentage finished sequence of the totalapproximately 3 Gb. Numbers are taken from the Genome Monitoring Table (Beck, 1998)(http://www.ebi.ac.uk/~sterk/genome-MOT/). * Estimated

MUTATION DETECTION

The spectrum of mutations ranges from large chromosomal rearrangements to small deletions,

insertions and single base substitutions. Identification of mutations in nucleic acids is an

important part of genetic research, not at least for the prediction and diagnosis of diseases.

Many methods for detecting mutations exists, based on a wide range of different technologies.

There is a constant flow of reports of new or improved methods, all with the hypothetical goal

of obtaining an optimal system, in the aspect of sensitivity, selectivity, simplicity, reduced time

and cost etc. Certainly, the complete determination of the whole DNA sequence in the region

of interest, is the optimal result that can be achieved. DNA sequencing is although still

relatively time consuming and costly and therefore a need for complementary methods exists.

These could at least be a first selective step to sort out a reduced amount of samples later

0

40

80

120

160

200

240

280

320

pre-1990 1990 1991 1992 1993 1994 1995 1996 1997 1998 (Sep)

1998(Dec)*

[Mb]

0

1

2

3

4

5

6

7

8

9

10

11

[%]

12

subjected to complete sequence determination or even obviate the need for sequencing if the

method allows for description of the exact nature of the mutation.

The different methods can be divided by their purposes, scanning or screening. Scanning

procedures are aiming at the search for unidentified variations, in comparison to the known

wild type sequence, across one or several whole genes or parts thereof. Screening methods are

on the other hand more devoted to detecting predefined mutations at specific positions, so

called hot spots. Those should be able to handle larger amounts of samples, since population

based carrier screening can be required for efficient identification of disease alleles. The

scanning and screening terminology have been defined in different ways, such as that screening

methods is used for detection of unknown mutations (here called scanning) and diagnostic

methods is used for the detection of defined mutations (here called screening) (Cotton, 1993).

Some of the most commonly used methods for the detection of small alterations in a sequence

have here been selected for a brief overview of their basic original principles. All of them have

continuously undergone developments and there are many different variants existing of each

type here described, where each usually have been denoted with a new acronym.

Scanning for unknown mutations

The first methods to be introduced for routine analysis were all based on mobility shifts in gel

electrophoresis and only with the ability to determine if a sample contained sequence

differences, compared to a known wild type sequence, not what kind of changes that had

occurred. In the SSCP method (single strand conformation polymorphism analysis) (Orita,

1989), PCR products are heat denatured and rapidly cooled for avoidance of reassociation.

The single strands form sequence dependent conformations that influence gel mobility and

different mobility indicate sequence differences. Despite the fact that the sensitivity and

specificity is relatively low (Jordanova, 1997), SSCP is still widely used. Once the protocol has

been optimized for a specific analysis, the simplicity is attractive for routine analysis as a first

sorting step.

Denaturing gradient gel electrophoresis (DGGE) (Fischer, 1983) is a similar method in which

the differences in stability for double-stranded DNA of different compositions are utilized. The

point of denaturation where the double-strands starts to turn into single-strands, due to a

gradient in temperature or chemical denaturant, is detected by gel electrophoresis and single

nucleotide alterations are possible to detect.

13

A third method based on sequence dependent conformation variations is heteroduplex analysis

(HA) (White, 1992). PCR amplified sample and wild type DNA is mixed, denatured and

renatured, resulting in homoduplexes and heteroduplexes. If the heteroduplexes contain

different alleles, the presence of mismatched basepairs can affect the mobility through a gel.

Several different compounds have been used for the cleavage of mismatched regions in

heteroduplex DNA, usually analysed by the appearance of short fragments in a gel

electrophoresis, although with various degree of success. In chemical mismatch cleavage

(CMC) (Cotton, 1988) the mismatches are modified by osmium tetroxide or hydroxylamin and

then cleaved by piperidine. Enzymatic mismatch cleavage (EMC) methods involve RNase

(Myers, 1985), T4 Endonuclease VII (Youil, 1995) and the E. coli mismatch repair enzymes

(MutH, MutL and MutS) (Smith, 1996). The binding of MutS, which is responsible for the

recognition of the mismatch, has also been used alone without successive cleavage (Bellanne-

Chantelot, 1997).

Screening for known mutations

Knowledge of the position and the surrounding sequence of the eventual mutation investigated

allows for a discrimination of samples by differential sequence specific hybridization analysis.

This technique is broadly used in several of the existing screening methods, of which some also

have the possibility to determine not only the presence, but also the identity of the mutation.

The use of allele-specific oligonucleotides (ASO) for mutation detection was the first method

devoted for mutation detection (Cotton, 1993). Utilizing the principles of Southern blotting,

Wallace and coworkers (1981) could discriminate between the hybridizations of mutant and

wild type specific oligonucleotides to membrane immobilized mutant genomic material. A

direct immobilization of PCR amplified sequences on a membrane, followed by a measurement

of the relative quantities of a label from hybridized mutant or wild type oligonucleotides have

been denoted the dot blot format (Saiki, 1986). A reverse dot-blot format is then obviously

consisting of immobilized oligonucleotide probes to which an amplified target is hybridized

(Saiki, 1989). This format allows for simultaneous analysis of many different alleles. Numerous

copies of a single PCR product are let to hybridize under stringent conditions to an array of

immobilized oligonucleotides and after rigorous washing the positions of the labeled target can

be visualized.

14

By designing a PCR primer so that the 3'-end is localized on the position of the mutation, a

mismatch in that position will prevent extension and hence no amplification can occur. This

method was in 1989 described by several groups, extending the list of existing method

acronyms; allele-specific amplification (ASA) (Okayama, 1989), amplification refractory

mutation system (ARMS) (Newton, 1989), allele specific polymerase chain reaction (ASPCR)

(Wu, 1989) and PCR amplification of specific alleles (PASA) (Sommer, 1989).

Figure 2. A schematic view of the principles of some of the most commonly used scanning and screeningmethods for mutation detection. Adapted from Cotton, 1993.

Several variants exist of methods based on the mismatch discrimination displayed by DNA

ligase upon the covalent joining of adjacently hybridized oligonucleotides. If an internal end-

nucleotide is affected by a mismatch, the ligation is inhibited. The detection of the ligation

product was in the original method, oligonucleotide ligation assay (OLA) (Landegren, 1988),

performed with solid phase capture of one of the oligonucleotides via the biotin-streptavidin

system and the presence of a ligation product could be detected by a label on the second

oligonucleotide. With the avoidance of electrophoretic separation, the method has also been

shown to be amenable for automation (Nickerson, 1990). The discovery of a thermostable

SSCP

DGGE

HA

CMCEMC

ASO

ASA

OLA

PEX

Normal Mutant

Changedmobilityin gel

Cleavage

No duplex

No amplification

No ligation

No extension

15

ligase led to the introduction of ligase chain reaction (LCR) (Barany, 1991), in which the

ligated product of oligonucleotide pairs is exponentially amplified, if there are no mismatches

in the position of nucleotide joining.

Single nucleotide primer extension (PEX) or DNA minisequencing (Syvänen, 1990) is a

method where the identity of a mutation can be determined. A primer is annealed adjacent to

the site of the mutation on an immobilized target and is extended by DNA polymerase with a

single labeled nucleotide, if it is complementary to the site of mutation. By dividing the target

and primer into four different reactions, a different nucleotide can be added to each. Not only

the identity is determined but also the ratio between different alleles. This concept has also

been applied on oligonucleotide arrays (Pastinen, 1997).

A relatively recently developed technology for sequence determination of short DNA fragment

is called pyrosequencing (Ronaghi, 1998), where an approach radically different from all other

existing methods have been applied. A four-enzyme mixture is utilized for the detection of

released pyrophosphate groups (PPi) during DNA polymerase dependent extension.

Nucleotides are added to a primer/template duplex in an iterative way and eventual

incorporation on the primer is measured by the conversion of pyrophosphate, via ATP, to light.

DNA arrays

Sequencing by hybridization

The idea of sequencing by hybridization (SBH) is elegant and large efforts have been made

during the ten years that has passed since the first reports, to make this principle working not

only in theory but also in practice. It is not hard to imagine that the influence from the allele-

specific hybridization experiments described above was important for the appearance of the

SBH strategy. With the insight that a large amount of different oligonucleotides could be

separately immobilized on a solid support, it was realized that a target DNA sequence could be

determined if all possible combinations of a short oligonucleotide were orderly immobilized,

and to which the labeled target to be sequenced was hybridized. The sequencing strategy is

based on computer assisted assembly of the target sequence with the knowledge of the

sequence identity of the oligonucleotides that the target has hybridized to. For the step going

from a theory to a method that can be applied in practice is it of course necessary to find

hybridization conditions able to discriminate between match and mismatch situations.

16

Figure 3. A schematic outline of the principle of sequencing by hybridization. All possible combinations of ashort oligonucleotide, which is here exemplified with dimers, are immobilized on a solid support. The target,which is here a hexamer, is hybridized under conditions where only fully complementary sequences can createstable duplexes. The immobilized complementary oligonucleotides are identified by a label on the target andthe overlapping sequences can be assembled to a continuous complementary sequence, revealing the targetsequence.

Two groups independently described the theoretical principle of the methodology in papers

submitted in 1988 (Lysov, 1988; Drmanac, 1989). Drmanac and coworkers envisioned a dot-

blot format (also denoted format 1) in which 95 000 11-mers consecutively should be

hybridized to immobilized genomic clones in order to sequence 1 Mb. In contrast, Lysov and

coworkers foresee a reverse dot-blot format (format 2) with a complete set of immobilized

octamers (48=~65 000) for the sequence determination of a target of a few hundred basepairs.

However, the experimental results of sequencing by hybridization did not indicate an

immediate success and most involved researchers had finally to realize that SBH was not

perfectly suited for de novo sequencing, but definitely for other types of sequence-based DNA

analysis, such as confirmatory sequencing and mutation detection (Drmanac, 1998).

DNA chips

AA CTAG TTCT TA CTTAAG GAATTCTA AATT AG

GAATTC GAATTC AA AC AG AT

GAATTC CA CC CG CT

GA GC GG GT GAATTC GAATTC TA TC TG TT

17

The introduction of chip based technology for the creation of high-density microarrays of

oligonucleotides has recently had an enormous impact on sequence-based DNA analysis. It

was originally the need to develop functional SBH array systems that led to the development of

the now existing chip technology. The procedure can be summarized in five steps; i)

immobilization of oligonucleotides or amplified DNA fragment on the chip which usually is a

coated glass plate; ii) sample preparation including labeling and fragmentation; iii)

hybridization and washing; iv) scanning; v) complex algorithms for data interpretation.

Several groups have developed variants of methods for immobilization of nucleic acids. The

oligonucleotides can either be attached presynthesized or synthesized in situ, on the chip, as

described e.g. by Southern et al (1992) who used standard reagents for oligonucleotide

synthesis in a circular reaction chamber. They have applied a combinatorial technique for the

arraying of all possible matching oligonucleotides in a given region for the selection of

antisense reagents (Milner, 1997). Another technique for in situ synthesis is the

photolithography (Pease, 1994), which is adapted from the semiconductor industry and based

on light-directed removal of protecting groups from the nucleotide with the aid of a masking

system. All possible oligonucleotides of a given length (n) can be synthesized in 4 x n cycles.

The current generation of technology allows for the fabrication of approximately 320 000

distinct oligonucleotides in a 1.6 cm2 array, where each spot is 20 µm in diameter and contains

~107 probes (Wang, 1998). Such chips have recently been successfully applied to extensive

analysis of sequence variations (see e.g. Chee, 1996; Hacia, 1996; Kozal, 1996; Wang, 1998;

Winzeler, 1998) and monitoring of gene expression levels (Lockhart, 1996; Wodicka, 1997).

Mirzabekov and coworkers have developed another variant of attaching pre-synthesized

oligonucleotides to the chip surface, namely via patches of activated polyacrylamid in small gel

pads (Yershov, 1996). This technique has also been used for immobilization of PCR products

(Proudnikov, 1998). However, this is not a high-density array system, recent published reports

with analysis of temperature melting curves for mutation detection (Drobyshev, 1997) and

thermodynamic analysis (Fotin, 1998), demonstrated the use of 64 pads on a chip. The

immobilization capacity for each spot is although higher than for other systems.

Researchers at the Stanford University, CA, have developed an array system where amplified

DNA is immobilized on the chip instead of oligonucleotides. They have for instance PCR

amplified all open reading frames (ORF) from the yeast genome and spotted those more than

6000 amplicons on the arrays, which have been used for genome wide comparative gene

expression analysis (DeRisi, 1997; Lashkari, 1997a; b). Furthermore, oligonucleotides

18

corresponding to all genes in the yeast genome have been arrayed on a chip by light-directed

photolitographic synthesis (as described above) for analysis of genetic selections and genome

wide transcriptional analysis (Cho, 1998a; b). For a recent review on the advances in

microarray technology, see Schena, 1998.

The sample preparation is a crucial step. Fragmentation of the amplified target can be utilized

to obtain less secondary structure and lower risk for multiple interactions. Different means of

fragmentation have been described, such as cleavage by deoxyribonuclease (Wang, 1998) or by

fragmentation of RNA. In the latter case, the target DNA amplification or the reverse

transcription of mRNA to cDNA is performed with primers containing RNA polymerase

promotor sequences, for subsequent in vitro transcription. The instability of RNA is utilized for

heat and salt fragmentation to 20-40 mers (Drobyshev, 1997) or 30-50 mers (Wodicka, 1997).

The labeling procedure is also an important part of the sample preparation and several different

variants have been explored (Chee, 1996). In general, the last step of nucleic acid amplification

is utilized for incorporation of dye-labeled primers or polymerase assisted incorporation of

fluorescent labeled nucleotides for direct scanning after hybridization or with biotinylated

nucleotides, which require a post-hybridizational adding of a dye-streptavidin conjugate.

All those described chip technologies are depending on having some sort of label incorporated

into the sample. One way to simplify the DNA array technology is to remove the need to label

hybridized samples. Alternative methods for detecting the specific spots where stable duplexes

have been formed must then be found. Since the hybridization signals often are spread over the

whole spectrum of signal intensities, due to the problems to discriminate all perfect matched

hybridizations from all mismatched situations, a system which at least is capable of measuring

relative quantities of hybridized sample DNA is desired. The label-free detection principles

developed for biosensors could possibly be imagined to be applicable also to high-density DNA

micro arrays.

BIOSENSORS

Generally, biosensors can be regarded as real-time analytical devices comprising a transducer

coupled to a sensing unit capable of detecting biological molecules at a surface. Such

molecular sensors combine the specificity of a biological recognition mechanism with a

physical transduction technique, which can be based either on electrochemical, mechanical,

19

thermal or optical sensing principles. The most usual aim of a biosensor is to produce

continuous signals that are proportional to the amount of molecules bound to or reacting at the

sensor surface.

There are a wide variety of transducers used in biosensors and a brief overview of the different

types of sensing principles will be described here, with the focus on recent developments in

nucleic acid based biosensors (Bier, 1997a; Yang, 1997; Zhai, 1997). Genosensors, which they

also have been denoted, have gained large attention as attractive analytical tool as late as in the

beginning of the 1990’s, whereas the more traditional enzyme sensors, immunosensors and

microbial sensor have undergone several decades of development. Nucleic acid based

biosensors normally employ immobilized DNA oligonucleotides or fragments on the sensor

surface, as the recognition element. Sequence specific hybridizations or interactions with

DNA-binding proteins can then be monitored and analyzed.

Electrochemical transducers

Amperometric devices

The measurement of electric current flowing at constant potential is the principle behind

amperometric biosensors, which typically utilize electrodes coated with enzymes which enables

the analyte to undergo oxidation or reduction, resulting in a detectable current. The magnitude

of the catalysis generated current has been shown to be proportional to the substrate

concentration (Lorenzo, 1998). Although, the majority of commercial biosensor devices are

amperometric enzyme electrodes, they have not made a deeper impact on the markets where

continuous monitoring is required (Griffiths, 1993).

Each amperometric biosensor is designed for detection of only one or maybe a few molecular

species, depending on the specificity of the enzyme. The definitely most commonly used

biosensor is the single-use amperometric glucose oxidase based sensor for home glucose

assays by diabetic patients. Another common type is sensors for hydrogen peroxide detection

in flow systems. Immunoassays can also be performed with amperometric detection, where

secondary antibodies are labeled with an enzyme whose product is oxidized at the electrode

(Heller, 1996).

Historically, major problems have been concerning how to functionalize the electrode surface

for site-specific immobilization of DNA fragments (Moser, 1997). Several examples of

electrochemical sensors for detecting DNA hybridizations have nevertheless been described

20

(Mikkelsen, 1996). The most common amperometric strategies for detecting the hybridization

of unlabeled DNA strands rely on redox-active hybridization indicators that bind more strongly

to DNA duplexes than to single stranded DNA, such as a minor groove binder (Hashimoto,

1994) or intercalating metal complexes of e.g. ruthenium or cobalt (Millan, 1994). Recently it

was showed by Napier et al (1997) how electron transfer could be monitored by cyclic

voltammetry for the specific detection of PCR amplified DNA with increased discrimination

between single stranded and hybridized DNA. Guanines have an enhanced reactivity compared

to the other nucleotides (Thorp, 1998) and the electron transfer is originating from oxidation

of guanines, catalyzed by a redox-active metal complex, such as ruthenium bipyridine. An

oligonucleotide probe with all guanines substituted for inosines is immobilized and let to

hybridize to an unlabeled guanine-containing target and a catalytic current is detected upon

hybridization to the immobilized probe. Despite that the introduction of inosines by definition

is leading to decreased sequence specificity, unpurified PCR amplicons of different origins and

lengths could be specifically detected.

Potentiometric devices

Another closely related group of electrochemical transducers are the potentiometric devices,

where potential changes is measured at constant current (usually zero). A logarithmic

relationship exists between the potential generated at the electrode surface and the activity of

the ion of interest. This leads to a wide dynamic range (Griffiths, 1993). Potentiometric

enzyme electrodes were the first to be investigated for biosensing purposes and the first

description of a biosensor, where the bio-component was an enzyme operating in combination

with an electrochemical transducer came in 1962 (Clark). Glucose oxidase was entrapped at an

oxygen electrode and the measured oxygen concentration was proportional to the glucose

concentration. Typically, an ion-selective or gas-sensing electrode is used to monitor changes

in product or substrate concentrations within the enzyme layer by detection of the presence of

the charge of an adsorbed analyte. The range of substrates to analyze is rather limited and the

gas sensors (CO2 and NH3) are sensitive for background variations and the ion-selective

electrodes are associated with problems of selectivity.

During the last years, the development of potentiometric methods for biosensing purposes has

taken significant steps forward, not at least for analysis of nucleic acid hybridizations.

Sequence-specific detection have been reported for hybridizations between 21-mers and 42-

mers (Wang, 1996). Another recent example of the advancement is described by Wang et al

21

(1997b) where mismatch discrimination was obtained for hybridizations between 17-mers.

PNA (peptide nucleic acid, recently reviewed by Eriksson, 1996 and Corey, 1997) and DNA

probes were immobilized on carbon paste electrodes and oligonucleotides with or without a

centrally positioned mismatch were let to hybridize. A cobalt complex bound to the duplex was

used as a redox-active indicator for presence detection. The PNA probe was shown to give a

significantly higher degree of discrimination under the chosen conditions, but with a five times

higher detection limit, which was claimed to be 0.2 µM for the complementary 17-mer and

even higher for a 30-mer.

A light-addressable potentiometric sensor (LAPS) can detect changes in pH at a silicon chip

surface and with a rather complex set-up, PCR amplicons (Olson, 1991), plasmid fragments

and cDNA (Dill, 1997) have been specifically detected. Probes labeled with either biotin or

fluorescein are mixed and let to hybridize to the target, and subsequently, streptavidin is added.

The formed complex is captured on a biotin coated nitrocellulose membrane and an

urease/anti-fluorescein antibody conjugate is thereafter bound. To create an assay

microvolume, the membrane is pressed against a silicon chip in the presence of urea. The

bound conjugate catalyzes the production of ammonia and carbon dioxide, which leads to a

detectable change in pH. The signal is displayed in µV/s and is said to be proportional to the

amount of analyte, i.e. the nucleic acid target. However, the complex set-up includes several

binding reactions that have to be fully efficient in order to fulfil that criterion.

Conductimetric devices

Conductimetric detection principles have also been used to some extent for biosensor

applications. The solution conductivity is changed when an enzyme reaction generates a net

change in the concentration of some ionized species. Recently, a biosensor that uses a mimic of

biological sensory functions i.e. ion-channel switches (ICS), have been described (Cornell,

1997) as very flexible and sensitive and suitable for a wide range of applications. The sensor is

essentially an impedance element and is based on that the conductance of a population of

molecular ion channels in a lipid membrane is switched by the recognition event.

Mechanical transducers

Piezoelectric devices

22

When an electric field is applied across a so-called piezoelectric material, such as quartz

crystal, it induces a mechanical stress in the material and creates an oscillating motion of a

certain vibrational resonant frequency. The frequency is depending on the mass on the surface

of the oscillator and decreases upon increased mass on the surface. The piezoelectric effect can

be utilized for analysis of a wide range of mass measurements, including biological recognition

events. Changes of less than 1 ng/cm2 can be detected (Ward, 1990). The mass sensing format

is commonly referred to as quartz crystal microbalance (QCM), alternative electrochemical

coupling results in other piezoelectric vibrational modes, such as a surface acoustic wave

(SAW). This format is also called piezoacoustic, where also the amplitude of the mechanical

wave is analyzed for the dependency of the mass changes at the surface. Such piezoelectric

devices have earlier mainly been used for measurements of mass changes in the vapor phase,

but both QCM and SAW have also been applied to analysis in liquids, with e.g. enzymes or

antibodies immobilized as recognition element (Minunni, 1994; Storri, 1998)

Despite the attractive label-free mass sensitive detection principle, experiments involving

detection of nucleic acids have been shown relatively little attendance during the many years of

development of piezoelectric biosensors. The relation between the mass and frequency was

described already in 1959 (Sauerbrey) and hybridization sensors in general, were actually first

discussed in combination with a piezoelectric transducer (Fawcett, 1988). Progress have

nevertheless been obtained in the last few years and some reports on detection of specific

nucleic acid sequences have been published (Ito, 1996; Nicolini, 1997), both involving DNA

fragments of around 4 kb. Su et al (1995) have described attempts of kinetic determinations

for hybridization between 4 kb fragments, where the classical Sauerbrey expression of the

relation between frequency and mass however had to be correlated after measurements of

incorporated radioactivity.

Recently, substantial progress in applications for the analysis of DNA hybridization has been

described. Time resolved kinetic measurements of hybridization between oligonucleotides with

a QCM, with a high frequency device for increased sensitivity, have been performed by

Okahata et al (1998). The immobilization was performed using the avidin-biotin system.

Through calibration of the QCM, the negative effects of interfacial liquid properties (i.e.

density, viscosity, electrode morphology etc) could be avoided. However, for longer DNA

fragments (>100 nt) it was not possible to obtain a linear correlation between the accumulated

mass and frequency. Interestingly, mismatch discrimination was obtained for hybridizations of

10-mer and 20-mer pairs, respectively, involving either a perfect match or one centrally

23

positioned mismatch. The association- and dissociation rate constants could be determined and

the effect on hybridization of spacer length, ionic strength and temperature was also

investigated.

A self-assembling monolayer of thiol-derivatized PNA on a gold plated QCM was the key

component in experiments described by Wang et al (1997a). Upon addition of a

complementary DNA oligonucleotide to the immobilized PNA (both 15-mers), the frequency

decreased, but remained unchanged when a 15-mer with a centrally positioned mismatch was

added at five times higher concentration. The detection limit for the complementary

oligonucleotide was 1 µg/ml, corresponding approximately to 0.2 µM for a 15-mer.

Thermal transducers

Calorimetric devices

Calorimetric biosensors, such as the enzyme thermistor (Danielsson, 1991), utilize that the

conversion of substrate to product generally is a more or less an exothermic process. The

enzyme is immobilized on a thermostatted column and what is actually monitored is the

temperature dependent resistance of the column. The biologically generated temperature

increase is proportional to the substrate concentration and the sensitivity is in the 10-50 moC

range, which have been shown to correspond to substrate concentrations in the lower µM

range (Danielsson, 1991). Assays have so far included detection of a wide variety of

compounds such as glucose, urea, ethanol, penicillin and sucrose.

A calorimetric device, which not can be regarded as a biosensor, but can be used for the

analysis of binding affinities between biological molecules, without the need to modify or have

one interacting molecule specie immobilized is the isothermal titration calorimetry (ITC)

device. The ITC is regarded as one of the most accurate methods for determining binding

affinities (Doyle, 1997). Since the binding enthalpy change is measured, i.e. the absorbed or

released heat during a binding reaction, the thermodynamic characteristics can also be

revealed. Analysis described involving nucleic acids have mainly been with investigation of the

action of DNA binding proteins (see e.g. Lundbäck, 1998; Plum, 1995). However, ITC have

recently been applied for thermodynamic analysis of DNA/DNA, PNA/DNA and PNA/PNA

decamer hybridizations (Ratilainen, 1998). Differential scanning calorimetry is a methodology

related to ITC and Breslauer and coworkers reported in 1986 how this method was used to

predict DNA duplex stability from the base sequence. This work resulted in an algorithm for

24

prediction of melting temperature for oligonucleotide duplexes, the nearest-neighbor method,

which is now widely used.

Optical transducers

Evanescent wave based devices

The reflectance of light at a surface is altered by the presence of surface bound analytes. The

common principle for optical biosensors is how the light is behaving at boundaries between

media of different refractive indices. Fundamental properties, such as intensity, phase and

angle, of the reflected light is measured directly and used to calculate the mass of analyte

molecules at the solid/liquid interface. The refractive index at the interface is changed in

proportion to mass changes at the surface, i.e. upon binding of molecules. In order to clarify

the relations between an evanescent wave, changes in refractive index and the measured signal,

a summary will follow of the techniques used in the commercially available optical biosensors

for real-time monitoring of label-free biomolecular interactions.

Surface plasmon resonance

The existence of surface plasmons had been anticipated for a long time before they were first

experimentally described in 1959 (Turbadar). Surface plasmons are charge density waves,

caused by collective excitation of free oscillating electrons. These occur at and propagates

along the interface between a thin layer of metal and a dielectricum, which in this context is an

aqueous solution (Liedberg, 1993). Surface plasmons can not be directly excited by light,

instead is a so-called evanescent wave used for indirect excitation (Lundström, 1994).

When the incident angle of a light beam propagating through an interface between higher and

lower refractive indices, exceeds a specific critical angle, total internal reflection (TIR) will

occur. In 1971 it was described how to excite plasmons in metal films under attenuated total

reflection (ATR) conditions, by directing the incoming light through a glass prism with direct

contact to a glass slide onto which the metal film is attached (Kretschmann, 1971). At total

reflection there is still a small portion of the light not being reflected, and this part of the light

constitutes the evanescent wave; an electromagnetic field that passes from the interface into

the lower refractive index medium. The evanescent wave interacts with the electrons in the

metal film and sets up a surface plasmon.

25

When the energy and momentum of the evanescent wave coincide with the surface plasmon

charge density wave, the phenomenon of surface plasmon resonance (SPR) can be observed as

a sharp dip in the intensity of the reflected light at a specific angle, the resonance angle. The

dip corresponds to a loss in energy in the reflected light, which is created by the transfer of the

photon energy to the charge density wave. The light energy is converted to heat. SPR occurs

in so-called free-electron metals where the electrons exhibit a gas-like behaviour, such as in

aluminium, copper, silver and gold. Another requirement is that the thickness of the thin metal

film is only a fraction of the wavelength, which should be monochromatic and in the visible

region. For a thorough review on the theory and physical properties of surface plasmons see

(Welford, 1991).

The evanescent electric field outside the metal decays exponentially with the perpendicular

distance from the surface and a typical penetration depth for the evanescent wave is 100-200

nm (Garland, 1996), it is however directly depending on the wavelength of the incident light.

Changes of the optical properties in the evanescent field will alter the resonance angle, with a

larger influence the closer to the surface the changes occur. The relation between changes in

optical properties and resonance angle is the basis for the use of SPR for biosensing purposes

and real-time biomolecular interaction analysis. Molecules that bind to the surface alter the

refractive index in proportion to the accumulated mass and the SPR angle is correspondingly

changed according to the mass of the bound molecules. Refractive indices measured for

specific wavelengths are convenient indications of the dielectric constants, which is the actual

physical parameter that influences the shift in the resonance angle.

In the BIACORE instruments (Biacore AB, Uppsala, Sweden) (Jönsson, 1991; 1992) the

resonance angle is called a resonance signal and expressed in resonance units (RU), since the

measured quantity is the interpolated position of the SPR-minimum and not the resonance

angle itself (Lundström, 1994). A linear correlation between resonance angle shift and surface

concentration has been demonstrated (Stenberg, 1991), with measurements of radioactively

labeled molecules and it was found that 1 RU approximately corresponds to a mass change of

1 pg/mm2, and an angle shift of 0.1 m° (millidegree). The detection limit is depending on the

stability of the biomolecular system, but is approximately 10 RU and since the detected area is

0.2 mm2 this is corresponding to 50 pg (Jönsson, 1991).

26

Figure 4. The principles of surface plasmon resonance (SPR). The evanescent wave causes the phenomenon ofsurface plasmon resonance. This is observed as minima in the intensity of the reflected light and the angle ofthis dip in energy, the resonance angle, is depending on the refractive index and the mass in the close vicinityof the sensor chip. The resonance angle is changed upon molecular binding and the shift in angle is convertedin a sensorgram to a plot of the shift in resonance signal versus time.

The first use of surface plasmon resonance for biosensing purposes was demonstrated in 1983

(Liedberg), where immunoglobulins were spontaneously adsorbed to a silver surface and the

subsequent binding of a second antibody was detected. SPR had earlier been used to

characterize organic molecular monolayers in gas phase (Pockrand, 1978).

Since the evanescent field is extended from the metal surface, a hydrogel matrix can be used

for biospecific interactions with one of the interacting molecule species covalently bound to the

matrix instead of direct adsorption to the metal surface. By introducing a dextran layer on the

gold surface (Löfås, 1990), a hydrophilic biocompatible surface and pseudohomogeneous

(semi-liquid) conditions can be achieved. Despite the extension out to the weaker evanescent

field, the sensitivity is retained and even increased due to the possibilities for higher amounts of

immobilized molecules in the evanescent field. An immobilization level corresponding to 1.5

monolayers in a matrix with a thickness of 100 nm have been shown to result in approximately

the same change in resonance angle as a monolayer adsorbed directly on the surface (Löfås,

Lightsource

Prism

Resonanceangle

Detectionunit

Sensorchip

Flowcell

I

II Resonancesignal

II

I Time

ReflectedIntensity

AngleI II

Sensorgram

Glass prism

Metal film

Evanescent wave

Incoming light Reflected light

27

1991). The first in the family of BIACORE instruments was introduced to the market in 1990

for real-time monitoring of biomolecular interaction analysis, where the acronym BIA has

become synonymous to the technology in general.

Figure 5. The sensor chip and the integrated micro-fluidics cartridge (IFC). The thin gold film is attached ontoa glass support, which is in contact with the prism. The hydrogel dextran matrix is immobilized via a linkerlayer. Four flowcells are created when the chip is docked on top of the IFC and the flow can then be directedwith pneumatic valves.

Besides the SPR based optical detection system, there are two important constituents of the

instrument, the integrated micro-fluidics cartridge (IFC) and the sensor chip. The IFC controls

delivery of buffers and samples, together with an autosampler, to the sensor chip surfaces and

channels are opened and closed through pneumatic microvalves (Sjölander, 1991). The sensor

chip consists of a glass support coated with a 50 nm gold film onto which the biocompatible

and flexible dextran layer is attached. The IFC is designed in such a way that when the sensor

chip is docked onto the IFC, four separate flowcells or reaction chambers are created.

Injections can be performed in series, with consecutive flow over all four flowcells, enabling

exact comparisons between the interaction of the injected analyte and different immobilized

ligands.

Another recently introduced SPR based instrument is IBIS (Instrument for Biomolecular

Interaction Sensing) (Windsor Scientific Ltd., Berkshire, UK / Intersens Instruments,

Amersfoort, The Netherlands). Consequently, the available amount of information regarding

the experimental performance of the instrument is very limited. The system is equipped with an

autosampler for injections and the biomolecular interactions occur either in a cuvette or in a

flow through system. They actually provide a one-channel adapter for BIACORE sensor chips.

Flow cellSensor chip

Valve

IFC

IFCD e x t r a n

G o ld f ilm

G la s s

Sensor chip

28

The quotation “Use the IBIS disks for simple routine assays and the more expensive

BIACORE sensor chips for publication material” follows the line of a low cost instrument,

which it is claimed to be. The IBIS instrument is displaying the response signal directly as m°.

Figure 6. Schematic representations of two waveguiding devices, a resonant mirror and a grating coupler. Ashift in the phase of the reflected light corresponds to a change in the refractive index in the evanescent field.

Optical waveguiding

Resonant mirror (RM) devices (Cush, 1993) are not depending on surface plasmon resonance,

but still utilize an evanescent wave to detect changes in the refractive index profile across a

solid-liquid interface. The sensing surface in this case is a dielectric resonant layer of high

refractive index, such as titania oxides or silicon nitrides. An incident beam of light is coupled

to the device by a prism. At certain angles of the incident light, where the phase is matching the

resonant mode of the high refractive index layer, the resonant point, light couples into the high

index layer, where it undergoes multiple total internal reflections and propagates along the

sensing interface before coupling back into the prism. This phenomenon is known as

Evanescentwave

Sample liquid

Incidentlight

Reflected light

Prism

Low indexcoupling layer

High indexresonant layer

Hydrogellayer

Resonant mirror

Grating coupler

Waveguidingfilm

Evanescentwave

Incidentlight

Reflectedlight

Grating

Sampleliquid

Glass slide

29

waveguiding. The angle of excitation of resonance is very sensitive to changes in refractive

index at the sensing interface. When the resonance angle is changed, a phase shift in the

reflected light can be detected as a peak in reflected light, instead as a dip in SPR, and the basic

principle behind the resonant mirror technique is the dependency of the phase shift on the

refractive index in the sensing layer (Hutchinson, 1995). A resonant mirror based instrument

was launched in 1993, the IAsys (Affinity Sensors, Cambridge, UK) (Cush, 1993). It works

with an open stirred and dextran coated cuvette system and has later been equipped with an

autosampler for a more automated operational mode.

The grating coupling instruments (Tiefenthaler, 1992) is based on a variation of the optical

evanescence technique, where a diffraction grating is used to couple the evanescent wave

instead of the low refractive index layer in the resonant mirror devices. Both do have a high

refractive index layer, as a waveguiding film (Ramsden, 1997). The IOS instrument (Integrated

optical sensor) (Artificial Sensing Instruments (ASI AG), Zurich, Switzerland), also launched

in 1993, constitutes of a more manual and also more modular system. The small-company

produced IOS device is hardly present at all on the list of published research reports based on

the different instruments, while there are Swiss and German academic groups working with

their own variants of waveguiding grating coupler systems utilizing modules from ASI

(Lukosz, 1995; Polzius, 1997).

Other optical devices

Devices based on total internal reflection fluorescence (TIRF) measure the binding of a

fluorescence-labeled ligand to its partner at the surface, which is recognized by the

fluorescence excited by the evanescent wave close to the surface. Non-specific adsorption of

fluorescence labeled molecules to the surface is the major practical limitation. Optical fibers

have relatively frequently been used as a solid support for TIRF analysis for sequence specific

detection of nucleic acids and even for direct synthesis of oligonucleotides on the fiber

(Piunno, 1995). Detection of single base mismatches in oligonucleotide targets have been

shown by (Kleinjung, 1997). A wave guiding alternative with a selenium-antibiotin conjugate

as a source of light scattering have also been demonstrated in a small chip format, for the same

purpose (Stimpson, 1995). An attempt to detect a point mutation in a 109-bp PCR amplicon

has also been reported (Healey, 1997), where two different oligonucleotides were immobilized

on optical fibers. In order to be able to discriminate between hybridization upon a 10-mer with

a mismatch and a perfect matched 11-mer, the differences in melting temperature between the

30

oligonucleotides had to be utilized by using a temperature just below the melting temperature

of the 11-mer. The PCR amplicon was biotinylated and detected by addition of a streptavidin-

fluorophor conjugate.

Imaging ellipsometry for visualization of biomolecular interactions, such as detection of

antibody-antigen complexes, is based on an optical biosensor device with a CCD camera for

analysis of surfaces. Ellipsometry is based on the change of polarization of light due to the

reflection from a surface. The change depends on the optical and microstructural properties of

the surface, which means for biosensing purposes the thickness of an adsorbed layer (Jin,

1995). The time resolution of an ellipsometer is relatively low, 1-10 s, (Ramsden, 1997)

compared to the millisecond range for the commercially available evanescent wave based

biosensors. Ellipsometry was the first method introduced to quantitatively investigate thin

layers at solid/liquid interfaces and it is mainly a surface characterizing technique, which can be

applied to all reflecting surfaces.

Flow cytometry is a method for measuring the fluorescence or light scatter of particles (Nolan,

1998). The analysis of molecular interactions can be made in real-time with subsecond kinetic

resolution, due to the possibility to discriminate between free and particle bound fluorescent

molecules. Fluorescent dyes associated with the particle are excited when sample flow passes

through a laser beam and the emitted fluorescence is detected. The particles have historically

mainly been cells for the analysis of ligand-receptor interactions on the cell surface. Recently,

also molecules have been immobilized on the surface of uniform microspheres and the second

molecule is fluorescently labeled and the interaction is measured as a change in the particle

fluorescence. Flow cytometry is also suitable for analysis of multimolecular complexes, with

different fluorophores on each interactant. Apart from kinetic studies of a wide range of

protein-protein interactions, DNA cleavage by endonuclease (Nolan, 1996) and the activity of

DNA ligase (Cai, 1997) have been measured in real-time.

Fluorescence correlation spectroscopy (FCS) is an interestingly method that have been applied

to kinetic analysis at single molecule level of an oligonucleotide hybridizing to a 7 kb target in

an extremely small volume element of 0.2 fl (Kinjo, 1995). FCS have also recently been applied

for single molecule detection of PCR products (Björling, 1998) and analysis of DNA

endonuclease activity (Kinjo, 1998).

31

RESULTS AND DISCUSSION

INTRODUCTION

The commercially available optical biosensors with capacity for real-time monitoring of

biomolecular interactions for determination of e.g. kinetic binding data (association and

dissociation rate constants) are now really established in many research areas. The two main

suppliers, Affinity Sensors and Biacore, have collectively more than 1100 articles in which

their instruments have been used, but almost 90% of those are based on the latter. In general,

the vast majority of analyses that have been performed can be categorized into the broad field

of protein-protein interactions. For an overview of applications, see e.g. Hutchinson, 1995;

Szabo, 1995; Malmqvist, 1997; Panayotou, 1998 and references therein. Nucleic acid analysis

have also been performed, although not at all to the same extent, employing nucleic acids

either as immobilized ligands for investigation of the action of DNA binding proteins or for the

analysis of interactions between nucleic acids.

The immobilization of nucleic acid templates on a sensor surface has been a major concern

during the years of research and development of nucleic acid based biosensors. Adsorption was

for long time the only alternative for immobilizing nucleic acids on the electrodes used as

sensor surfaces in electrochemical and piezoacoustic biosensor devices. It was a relatively

simple procedure with no requirements of surface or ligand modifications, but the random

immobilization led to low efficiencies in subsequent hybridizations. In addition, a wide range of

chemical modifications, via linkers introduced at the oligonucleotide and an adhesive agent on

the surface, have been used for site-directed covalent coupling (Yang, 1997). In 1990

(Ebersole) it was shown how an active monolayer of avidin could form on metal surfaces. The

streptavidin-biotin system is now a widely used procedure for the site-directed immobilization

of nucleic acids on a solid support, utilizing one of the strongest known non-covalent

biological interactions, KD ~10-15 M (Wilchek, 1988). The introduction of a flexible and

biocompatible dextran matrix (Löfås, 1990) in which streptavidin can be covalently bound,

have certainly increased the efficiency of nucleic acid immobilization. Watts et al (1995) have

compared different methods for immobilization of oligonucleotides on a sensor surface.

Adsorption and aminolinkage required almost 20 times higher concentrations to yield

maximum responses compared to the use of biotin-streptavidin capture in a dextran layer, but

the maximum response was still four times higher for the latter.

32

The possibilities for convenient and efficient immobilization of biotinylated nucleic acids

together with the mass sensitive detection system offered by the BIACORE instruments,

demonstrates an appealing technology for biomolecular interaction analysis. The herein

described experiments were initialized in order to investigate how far this surface plasmon

resonance based biosensor system could reach into the relatively untouched area of molecular

biology applications.

REAL TIME MONITORING OF DNA MANIPULATING ENZYMES (I)

To investigate if basal unit operations in molecular biology could be analyzed with optical

biosensor technology, a model system was designed, consisting of six oligonucleotides. They

could be assembled into a 69-bp double stranded DNA fragment and subsequently utilized as a

target for endonuclease cleavage and after strand separation, as a template for DNA

polymerases.

DNA assembly using DNA ligase

One oligonucleotide was biotinylated in the 5'-end for fast and efficient immobilization onto a

sensor chip where the dextran layer was coated with streptavidin. A double stranded fragment

with a 3-nt protrusion was achieved by letting a complementary oligonucleotide hybridize.

Two consecutive ligation steps with prehybridized oligonucleotides containing 3-nt protrusions

completed the assembly. The efficiencies in those non-optimized hybridization and ligation

steps were each calculated to approximately 64 % of the stoichiometric maximum. Control

injections without ligase or with non-complementary protrusions did not result in any

accumulated mass on the surface. Furthermore, the response curves for injection with ligase

and DNA fragment were not identical for the different ligation steps. A bi-phasic response

curve was observed for events where both strands were ligated, which was not observed when

only one strand was involved in the formation of a phosphodiester bond.

DNA assembly can be an alternative way of creating a desired DNA fragment with a length

longer than what is optimal for an ordinary DNA synthesizer. The possibility to monitor such

DNA assemblies in a step by step manner leads to the advantage of known efficiencies in each

step of the reaction. Such an assembly usually involve several reaction steps and there are often

no possibilities to evaluate the outcome until the process is finished and the end product is

analyzed, by electrophoresis or other means. The concept used here demonstrates not only the

33

opportunity to optimize reaction conditions, regarding buffer, temperature, ionic strength etc,

but also possibilities for basic research regarding the activity of DNA ligase. Besides being an

analytical tool, the methodology can also be used preparatively. Parts of the assembled DNA

fragment can either be eluted with cleavage by an endonuclease at a designed cleavage site.

Alternatively, the non-biotinylated strand could be eluted by chemical denaturation, and used

as template in subsequent PCR amplification to increase yields.

Endonuclease cleavage

The elution of the assembled fragment by endonuclease cleavage was also demonstrated, not

for preparative purposes but to evaluate if an endonuclease activity could be monitored. In

contrast to later released versions of the instrument, the model used here did not support

automatic recovery of eluted samples. The 69-bp DNA fragment contained the recognition

sequence for the endonuclease Xho I and injection of the restriction enzyme resulted in a clear

decrease of mass on the surface. The recognition site was positioned in such a way that

cleavage would leave a 16/20 nt long fragment on the chip. The net decrease obtained by

comparison of response levels before and after injection indicated a cleavage efficiency of 93%.

The real-time monitored response curve for the reaction is relatively complex to interpret,

since it represents the sum of three simultaneously occurring processes; enzyme binding,

enzyme dissociation and the release of DNA by the cleavage reaction. Nevertheless, the results

demonstrate the possibilities of kinetic analysis of the action of DNA endonucleases.

DNA synthesis by DNA polymerases

A third class of enzyme, which is very common in many types of molecular biology

applications, is DNA polymerases. Using the model system, the activities of T7 DNA

polymerase (T7) and the Klenow fragment of DNA polymerase I (Klenow) were investigated.

The biotinylated strand of an assembled 69 bp fragment served as a template and an annealed

primer was extended towards the biotin. Differences in the behaviour between the two

polymerases were evident just by a visual comparison of the obtained response curves. In

experiments using T7 a continuous increase in the response was seen during the whole

injection. In contrast, Klenow displayed a fast and narrow peak during the first part of the

injection and then reached a plateau value. Furthermore, the response levels after the injections

were twice as high for T7 compared to Klenow. Interestingly, after a short pulse of SDS for

34

removal of eventually accumulated enzyme on the DNA, the increases in response

corresponding to the DNA synthesis were almost identical. Interpretation of the results

suggests that T7 remains bound to the DNA fragment even after the extension is completed,

while the Klenow "jumps off" after the synthesis is finished. In other words, T7 display affinity

not only for partly double stranded DNA fragment (primer + template) but also for double

stranded DNA, which the Klenow enzyme does not. This simple assay should be useful also for

the characterization of other polymerases. Buckle and coworkers (1996) have also performed

real-time measurements of polymerase dependent elongation using a reverse transcriptase,

including quantitative analysis of the polymerization process.

Taken together, the work showed upon possibilities of using biosensor technology for

quantitative and qualitative analysis in real-time for the characterization and monitoring of the

action of several of the most commonly used DNA manipulating enzymes. Later, several other

reports of biosensor facilitated investigations of interactions involving such enzymes have been

reported. They include the kinetic and stoichiometric analysis of mutant variants of the E. coli

RNase HI binding to pre-hybridized and thereafter immobilized RNA-DNA hybrids of lengths

up to 36 bp. Also comparisons with the reverse set-up was performed, with the duplex injected

as analyte (Haruki, 1997). Another recent example of analysis of DNA manipulating enzymes

is the investigation of the mechanism of substrate recognition for inactive mutant variants of

the DNA repair enzyme uracil-DNA glycosylases described by Panayotou (1998). The IBIS

instrument has been used to characterize different mutant variants of the DNA unwinding

adenovirus single-stranded DNA binding protein by affinity determinations of the binding upon

an immobilized 41-mer oligonucleotide (Dekker, 1998).

MISMATCH DISCRIMINATION IN OLIGONUCLEOTIDES (I, II)

Minisequencing (I)

The results from the extension experiments described above implied the possibilities to develop

a strategy to employ the resulting SPR signal for a qualitative analysis of the template DNA

strand. Therefore, a minisequencing protocol based on biosensor analysis was designed and

investigated. Immobilized single stranded templates were generated by assembly and

subsequent strand separation, followed by primer annealing. DNA polymerase assisted

incorporation of one of the four dideoxynucleotides at the position adjacent to the primer in a

first step would prevent any further extension in a subsequent step using all four

35

deoxynucleotides. The identity of the base adjacent to the primer would then be revealed by

the absence of a signal related to the extension in the second step. Each experimental series

consisted of consecutive injections with i) alkali for strand separation, ii) primer annealing, iii)

polymerase together with one defined dideoxynucleotide, iv) SDS for removal of polymerase,

v) polymerase together with all four deoxynucleotides, vi) SDS. The response after step vi)

was compared for the four series with different chain-terminators. Thus, three high signals

were expected, corresponding to non-prevented extension and a fourth low signal,

corresponding to an incorporation of a dideoxynucleotide. The results showed, in principal,

that the technique worked, but to increase performance an alternative could be to utilize a four

flowcell system and have the template immobilized on all four surfaces and perform all

injections, except the one containing dideoxynucleotides, in a serial mode over all flowcells.

This would save time but increase the chip consumption. Another alternative for serial

injections would be to use four different primers for the analysis of four different mutational

hot spots.

An interesting feature of the minisequencing results can be added to the discussion regarding

the differences in DNA binding displayed by Klenow and T7 (see DNA polymerases above).

Klenow DNA polymerase was used in the minisequencing protocol and in the three cases of

extension, all obtained curves displayed a characteristic “Klenow extension shape” as observed

earlier. However, in the case with dideoxy-terminated extension a gradual increase was

observed during the injection. This further supports the theory that Klenow only show affinity

to partly double stranded templates, leading to an accumulation with time. The appearance of

this atypical curve also strengthens the discriminatory power of the minisequencing

methodology, by providing a second indication of which dideoxynucleotide that has been

incorporated. On the negative side of features displayed by Klenow is the 3'-5' exonuclease

activity, which could be a reason for the background extension signal that was observed. The

level of the background signal was 37% of the average net response observed for the other

three cases. The results suggest that the polymerase to use should be chosen with care.

Hybridization analysis by determination of equilibrium affinities (II)

In order to investigate the hybridizational behaviour of short oligonucleotides, a second

oligonucleotide based model system was designed. The rapid dissociation of a hybridized short

oligonucleotide from the template enables affinity determinations from a monitoring of steady-

state responses at equilibrium and it also circumvents the need for regeneration between

36

sample cycles. Affinity data are easily obtained from the plotting of equilibrium responses

versus the concentration. An analysis of the effect of temperature and oligonucleotide length

on affinity was performed for hepta-, octa- and nonamers hybridizing onto a sensorchip-

immobilized 25-mer. An approximate 10-fold increase in affinity was observed when the length

was increased from eight to nine nucleotides (in the temperature range of 20-35 °C), while the

difference between a heptamer and an octamer only were two-fold (10-35 °C).

The affinity determinations of eleven different octamers hybridizing to different positions on

the same complementary target revealed a difference in affinity of almost two orders of

magnitudes. A comparison to calculated Tm (melting temperature) values suggested that a

difference in Tm of 10 °C correspond to a 10-fold difference in affinity. Those results highlight

the fundamental difficulties encountered in early experiments designed for sequencing by

hybridization (see Introduction). It is obviously very hard just to find a single condition where

all perfectly matched hybridizations should result in stable duplexes, whereas mismatched

events should not.

Mutational scanning and influence of mismatch position (II)

The possibility to detect base substitutions in the target sequence through affinity

determinations of hybridizing oligonucleotides was also demonstrated. The sensitivity of the

strategy was shown by a comparison of the affinities between a nonamer and two different

targets, for which the oligonucleotide was either fully complementary or contained a single 5'-

end mismatch. The resulting affinity plots shows that even for an end-mismatch situation, the

discrimination factor is, expressed as difference in affinity, one order of a magnitude. After

such curves have been established, it is suitable to choose a single concentration for which the

largest differences are obtained. Thus, a subsequent discriminatory analysis can be performed

with only one concentration.

The effect of the relative localization of single mismatched bases was further analyzed using a

set of eleven octamers for scanning of three different 25-mers as targets, where two contained

a substituted nucleotide at different positions. By plotting the equilibrium responses for the

three targets a "mutational scanning profile" was obtained, which displayed how the mismatch

position dramatically influenced the responses. The effect was, as expected, more pronounced

when the mismatch was localized in an internal position. Noteworthy, a comparison of the

responses obtained for octamers hybridizing with a 3'- or 5'-end mismatch indicates that the

37

stability in this case was mostly affected by a 3'-end mismatch. This was not due to different

mispaired bases, since they were identical at both sides. The average response for the two

octamers with a mismatch in the 3'-end was only 18% of that obtained for hybridization to the

complementary target. The corresponding value for the 5'-end was 45%.

Figure 7. Overview of the different applications investigated in paper I-III, where oligonucleotide modelsystems were used.

The first description, in this context, of the monitoring or rather detection of oligonucleotide

hybridizations was submitted for publication (Wood, 1993) little more than a year after the first

launch of the instrument, demonstrating an early insight into the possibilities of molecular

biology application. More recent investigations of the influence of several different variants of

mismatches on the kinetic rate constants for oligonucleotide hybridizations have been reported

(Gotoh, 1995 and Jensen, 1997), where the latter work also included RNA and PNA

oligonucleotides. The ability for grating coupler devices to monitor and analyse DNA

hybridizations between oligonucleotides have been shown by Bier et al (1996; 1997b).

Mismatch discrimination was achieved through determination of kinetic data for 13-mers, with

B

DNApolymerase

DNA synthesis

B

ddCTP

Minisequencing

DNApolymerase

G

ddGTPddTTPddATP

DNA cleavage

B

DNAligase

Gene assembly

B

Immobilization

B

Hybridizations

B B

DNA endo-nuclease

B

Modular hybridizations

38

or without one or two centrally positioned mismatches, hybridizing upon an immobilized 24-

mer. Also the IAsys instrument has been utilized for real-time monitoring of oligonucleotide

(up to 40-mers) hybridizations. Investigations of the influence of oligonucleotide length and

the relative position of complementary sequence was performed (Watts, 1995). Furthermore,

antisense reagents have been evaluated by analysis of the binding of a triple helix forming

oligonucleotide and a DNA binding drug, separately or in combination, to an immobilized

oligonucleotide duplex (Rutigliano, 1998). Kinetic studies of oligonucleotide triplex formation

have also been described by Bates et al (1995).

Analysis of modular hybridizations (III, IV)

Increased stability by modular hybridizations (III)

Another type of hybridization analysis that has been performed was an investigation of the

stabilizing effect achieved upon co-hybridization of adjacently annealing short oligonucleotides,

the so-called modular primer effect. This effect can be utilized for the substitution of a single

longer primer or other oligonucleotides with a combination of shorter, e.g. hexamers.

Interactions between target DNA and pairwise combinations of both DNA and peptide nucleic

acid (PNA) hexamers were quantitatively analyzed for different adjacent hybridization

situations as well as overlap and one or two base gaps. The influence of concentration and

individual inherent affinities on the stabilizing effect was investigated in order to illustrate the

importance of those parameters for use in different modular hybridization applications, and

also to provide a suitable methodology with capacity for such analysis.

Again, the fast hybridization kinetics for short oligonucleotides, in this case hexamers, were

utilized for the determinations of apparent duplex affinities from equilibrium responses. Two

out of the four flowcells were utilized for the separate immobilization of two different but

closely related 25-mers, designed to result in different hybridization situations. In the third

flowcell, a mixture of two targets was immobilized, each separately comprising either of the

two annealing sites. Hence, any co-operative stabilizing was ruled out, allowing for

comparative analysis between separate and modular hybridizations.

Firstly, the modular primer effect was investigated at fixed total hexamer concentration, but

with different concentration ratios between the module pairs. Four different hexamer

sequences, where two were used both as DNA and PNA hexamers, were analyzed for seven

different combinations of pairwise injections. In order to observe co-operative effects, the

39

arithmetic sums of hybridization responses for corresponding hexamer combinations and

concentrations, previously obtained from injections of individual hexamers, were subtracted

from the responses obtained from the co-injections. Hence, any co-operative effects would

result in positive responses in the processed data. For situations of adjacent DNA hexamer

hybridizations, a clear net effect from the co-injection of modules was observed, demonstrating

an increased apparent affinity as compared to responses from separate injections of the same

modules. The inherent affinities of each hexamer had previously been determined and the data

shows that a more pronounced stabilization occurs after addition of small amounts of a higher

affinity hexamer module, as compared to the stabilization obtained after addition of small

amounts of a module with lower inherent affinity. Also for DNA hybridization formats

involving either a one-base overlap or two out of the three one-base gap situations, a small

modular stabilization effect could be observed. When DNA hexamer modules were replaced by

their PNA counterparts in different module combinations, it was found that also this nucleic

acid analogue with a different molecular backbone could contribute to a stabilization, either in

combination with a DNA module or as a PNA- PNA pair.

To further investigate the modular primer effect, a comparison was performed of apparent

combined affinities for two DNA hexamers with different inherent affinities, one injected

without or with different concentrations of the second module. Several injection series with

varied concentration of one hexamer and fixed concentration of the second hexamer

throughout each injection series, and vice versa, were performed to analyze the stabilization

dependency of module concentration. Large differences in combined apparent affinities

between hybridizations without any stabilizing module and with different degrees of adjacent

supplementation were observed. Up to an 80-fold increase in apparent combined affinity was

observed when a "low affinity" module was supplemented with a "high affinity" module and up

to a 20-fold increase was seen for the opposite combination. Also, for a hybridization situation

with a one-base gap, a small stabilizing effect could be observed. The stabilization factor was

approximately 5% of the value obtained by adjacent hybridization.

The use of oligonucleotide modules for capture (IV)

The favourable stabilization effect obtained by modular hybridizations was further utilized and

analyzed for the enhanced capture of single stranded PCR amplified DNA. The PCR amplified

target was a 324-bp fragment, originating from the nontranslated region of the hepatitis C virus

40

genome. The effects of different capturing probes on DNA hybridization, used together with

single oligonucleotides or multiple short modules prehybridized to the single-stranded target,

were investigated. The capturing probes were immobilized on the sensor chip and the PCR

fragment was rendered single stranded by streptavidin - biotin immobilization on a solid phase

followed by strand specific elution of the non-biotinylated strand. Capturing probes of 9 - 36 nt

were surprisingly not able to provide any significant capture of the single stranded DNA

fragment. However, when a prehybridizing step was introduced in which one or two

oligonucleotide modules were let to hybridize adjacently to the target sequence for the

capturing probe, a significant capture was observed. When an 18-mer capturing probe was

used together with single prehybridizing modules with lengths between 11 and 18 nt, all

combinations resulted in a high level of capture, while a prehybridizing nonamer only could

assist to capture a low amount of single stranded DNA. Interestingly, when the 18-mer capture

probe was shortened to a nonamer sequence and an extra prehybridizing nonamer (in total

two) was added, an effective capture was also obtained.

Figure 8. A cartoonlike view of the capture of single stranded DNA assisted by oligonucleotide modules.

In order to further analyze the oligonucleotide-assisted capture approach, single nucleotide

gaps were introduced between the modules and between the capture probe and the neighboring

module. All analyzed gap situations resulted in a significantly reduced capture, some abolished

the capture totally while some were not that detrimental with a 30% reduction. This gap-effect

was more pronounced for multiple modules.

Taken together, the results demonstrate that successful capture can be achieved using multiple

short modules and suggest that it is not primarily the length of the immobilized capturing probe

which is the most important parameter, but rather it depends on the number and length of

No capture Capture

27 nt9 + 9 + 9 nt

41

modules used in the prehybridization step. It can be envisioned that it would be feasible to use

presynthesized module libraries for nucleic acid capture by taking advantage of stacking

interactions between the oligonucleotides and possibly a co-operative suppression of target

secondary structure. The described biosensor technology can as mentioned earlier be used also

for preparative applications, which in this case could lead to an enrichment or purification of a

specific nucleic acid sequence. However, the use of the technology would probably mainly be

of analytic nature, to specifically optimize conditions of modular assisted capture of desired

target DNAs.

Mutation detection in clinical samples (V, VI)

Mutation detection on immobilized or injected PCR amplicons (V)

The introductory positive results on sequence-based DNA analysis with oligonucleotide based

model systems (I, II) were consequently followed up by investigations of the possibilities to

perform such analyses on clinically relevant samples, i.e. PCR amplified DNA.

Exon 6 of the human tumor suppressor p53 gene was chosen as target for the mutational

analysis using clinical samples derived from breast tumor biopsies. Two different hybridization

formats were utilized for mismatch discrimination between fully complementary sequences and

single base mismatches. The two formats, which can be recognized from the sequencing by

hybridization terminology, are format 1 with immobilized PCR products and oligonucleotides

injected for hybridizational comparison of targets and format 2, where oligonucleotides are

immobilized and single stranded PCR fragments injected. In format 1, two biotinylated PCR

products (199 bp) were immobilized, one of known wild type sequence and one with a 100%

single point mutation. After the non-biotinylated strand was chemically denatured and eluted, a

set of four different 17-mers were hybridized to the two targets. One oligonucleotide was a

positive control, which annealed with perfect match on both targets. Two of the

oligonucleotides were differing by only one base, one wild type specific and one mutant

specific. The fourth was also wild type specific, but the mismatch resulting from hybridization

to the mutant sample was localized closer to the end for this oligonucleotide. The results

demonstrated that relatively minor, but nevertheless reproducible and significant differences

could be obtained for hybridizations to the two targets. The two oligonucleotides differing by

one base resulted both in a small hybridizational advantage for their specific targets, while the

other wild type specific oligonucleotide possessed a higher discriminatory power.

42

Experiments in format 2 were initialized to obtain a methodology platform with possibilities to

screen many samples for known mutations, in contrast to format 1 which is more suitable for

the analysis of a longer sequence in few samples. The four flowcells now contained four

different oligonucleotides, a positive and a negative control and a wild type and mutant

specific. The molar capacity of the streptavidin surface was approximately eight times higher,

for the oligonucleotides compared to immobilizations of PCR products. The stoichiometry of

full-match hybridizations was on the other hand corresponding to a use of approximately 8%

of the available oligonucleotides, while in format 1 the corresponding value was approximately

50%.

The single stranded PCR products were generated with solid-phase technology as described

earlier. They were injected in series over the four flowcells and when the responses obtained

from the negative control surface was subtracted, a very clear difference could be observed

between the wild type and mutant sample. The responses from the capture by the specific

oligonucleotides of their complementary targets were gradually increasing during the whole

injection at a relatively high rate. Interestingly, the inverted combinations with a single base

mismatch resulted in no increase at all, leading to a clear mismatch discrimination where the

results in fact can be converted to the form of "1 or 0". A third sample containing a 50/50 mix

of wild type and mutant sequence was also analyzed and even in this case was the mutation

detected. A comparison of the hybridization signals from the wild type specific oligonucleotide

for the wild type sample and the mixed demonstrates a reduced hybridization response, but the

comparison has to be made also with respect to responses from the positive control surface.

Those experiments actually constituted the first published report, according to publicly

available databases, on hybridizational mismatch discrimination with PCR amplified DNA using

a label-free detection system. PCR amplified DNA has also been introduced into optical

biosensors by at least two other groups; Bianchi et al (1997) could specifically and direct

detect asymmetric PCR products amplified from HIV-1 genomic sequences and Gotoh et al

(1997a) have used the mismatch binding protein MutS with preliminary results of mismatch

discrimination between 95 bp homo- and heteroduplexes. The same PCR products were also

utilized for immobilization and hybridization with wild type specific and mutant specific 13-

mers and they obtained a weak mismatch discrimination (Gotoh, 1997b), in accordance with

the corresponding results shown here.

Mutation detection by subtractive hybridizational scanning (VI)

43

The above described methodology for mutation detection, despite the proven functionality, has

some drawbacks. The mismatch discrimination was under those conditions rather small in

format 1 and only three samples could be analyzed per chip, since one of these needs to be a

wild type reference. The limitations of format 2 lies in the relatively short region of the

sequence that can be analyzed, which although is enough for screening of mutational hot spots.

In order to overcome those limitations and to further develop the methodology, a totally new

concept was designed. In this study the mutation sensitive hybridizational scanning of PCR

product was performed outside the instrument, followed by an analytical hybridization analysis

between fully complementary oligonucleotides. The use of such subtracting hybridization on

the amplified target allows for a subsequent biosensor analysis of fast, robust and well

characterized interactions. Furthermore, the capacity in terms of number of samples and

sequence length to analyze is improved in comparison to when the PCR amplicons are

introduced in the instrument.

Sets of overlapping scanning oligonucleotides complementary to wild type sequence were

hybridized to microbead-immobilized PCR products. Mismatch hybridization situations

between a mutant sample and oligonucleotides result in higher remaining concentrations in

solution of involved oligonucleotides. Post-hybridization supernatants were subsequently

analyzed for their oligonucleotide compositions using the biosensor. Relative remaining

oligonucleotide concentrations were monitored through hybridization analysis between

scanning oligonucleotides and their corresponding sensor chip-immobilized complementary

counterparts. This allowed for the construction of composition diagrams revealing the

existence and approximate location of a mutation within an investigated sample DNA

sequence.

Again, the p53 gene was chosen as a model system and both exons 6 and 7 were scanned for

mutations. Fifteen 13-mers covering exon 6 and eleven 13-mers covering exon 7 were used for

scanning, all with six nucleotides overlap relative to flanking oligonucleotides. The scanning

oligonucleotides were divided into two non-overlapping sets for each exon to avoid

interference between oligonucleotides during hybridizations. Supernatants from each set and

sample were divided for injections onto two sensor chips. The supernatants from hybridizations

to sample and wild type reference PCR products were compared. Thus, a higher sample

derived hybridization response signal for a given oligonucleotide interaction reflects a higher

remaining oligonucleotide concentration in the supernatant, indicative of an earlier mismatch

situation due to mutation. The hybridizational responses from wild type supernatants were

44

subtracted from the responses obtained by sample supernatants. This processing of the data

enables the comparison to be performed without the need for equal PCR product

concentrations, since that will just be revealed as a baseline, i.e. the difference between

unaffected scanning oligonucleotides, with average values separated from zero. The complete

mutational scanning profiles can be regarded as a map of the exons, in which the processed

postinjection responses are plotted against the ordered identity of the scanning

oligonucleotides.

Figure 9. To increase the throughput of the method, a series of experiments were also performed withsimultaneous analysis of exon 6 and exon 7, involving the addition of two sets of scanning oligonucleotides to amixture of exon 6 and exon 7 derived PCR products and co-immobilization of two different oligonucleotidesper flowcell. A decreased sensitivity was observed, probably due to the increased complexity and the resultinglower immobilization level for each individual oligonucleotide species, although mutations were still detectableas exemplified here. The sample denoted K5L2 contains a mutation located in a sequence covered byoligonucleotides K and L.

The results presented in paper VI clearly demonstrated a reproducible detection of single point

mutations in five different clinical tumor samples. Due to the biological origin of these samples,

they were contaminated with various amounts of normal cells carrying wild type p53 alleles.

The present protocol used for subtractive hybridizational scanning is capable of detecting a

mutation in a background of approximately 50-60% wild type sequence.

CONCLUDING REMARKS

-50

-40

-30

-20

-10

0

10

20

30

40

50

60

A B C D E F G H I J K L M N O P Q R S T U V W X Y Z

Rel

ativ

e R

esp

on

se [

RU

]

Exon 6 Exon 7

K5L2

45

It has here been demonstrated how an optical biosensor, based on surface plasmon resonance

for mass sensitive detection and analysis of biomolecular interactions, can be utilized for a wide

range of molecular biology applications, with special emphasis on hybridization analysis and

mutation detection. The described biosensor system has been shown to be a suitable tool for

basic research related to investigations of kinetic properties of the hybridization event and with

capacity for discriminating between fully complementary hybridizations and hybridizations

involving a single-base mismatch. Several different methodologies for sequence-based analysis

of clinically relevant samples have here been described. However, without any knowledge of

eventual future instrument developments, it must be stated that the present instrumentation

with only four flowcells prevents the technology from high throughput analysis. The

automated, robust and sensitive system together with the convenient biotin/streptavidin

immobilization and most of all, the label-free detection principle is although an appealing

concept and it can be envisioned that the detection principle could be implemented to other

chip based system.

46

ABBREVIATIONS

ARMS amplification refractory mutation systemASA allele specific amplificationASO allele specific oligonucleotidesASPCR allele specific polymerase chain reactionATR attenuated total reflectionBIA biomolecular interaction analysisbp basepairsCA CaliforniaCCD charge coupled deviceCMC chemical mismatch cleavageDGGE denaturing gradient gel electrophoresisDOE Department of EnergyEMC enzymatic mismatch cleavageFCS fluorescence correlation spectroscopyGb billion basepairsHA heteroduplex analysisHGP Human Genome ProjectICS ion-channel switchesIFC integrated micro-fluidics cartridgeITC isothermal titration calorimetrykb thousand basepairsLAPS light addressable potentiometric sensorLCR ligase chain reactionMb million basepairsNADH nicotine adenine dinucleotideNIH National Institute of Healthnt nucleotidesOLA oligonucleotide ligation assayORF open reading frameOWLS optical waveguide lightmode spectroscopyPASA PCR amplification of specific allelesPCR polymerase chain reactionPEX primer extensionPNA peptide nucleic acidsQCM quartz crystal microbalanceRM resonant mirrorRU resonance unitsSAW surface acoustic waveSBH sequencing by hybridizationSDS sodium dodecyl sulphateSPR surface plasmon resonanceSSCP single strand conformation polymorphismTIR total internal reflectionTIRF total internal reflection fluorescence

47

ACKNOWLEDGEMENTS

I would like to express the most sincere thanks to all people that are working or have workedduring those years, in the Dept. of Biotechnology at KTH - Royal Institute of Technology andespecially in the fantastic and unique DNAcorner®. The people and atmosphere and also theintense activities in this place was the major component in my first thoughts of starting to workfor a PhD.

I would like to resemble the lab with a flock of sheep, where the chief shepherd is alwaysrunning around with a neverending enthusiasm and always with a happy tone in the horn,keeping an eye on the flock and constantly seeking for new ways of providing us occupied inmeans of bringing us to new green grazing lands. The herd is constantly growing, even if theland is not growing at the same speed, and all sheep were divided among the introducedsheepdogs, which take very good care of their groups, despite their burden of amount of work.The herd has an excellent mixture of individuals: the ones that try to keep the place tidy andorganized and looking after the black sheep of the herd; the ones that almost always have asmile in the face, even at tough situations; the ones that take their responsibilities seriously; theones that keep or bring things to work; the ones that always can give and take a joke; the onesthat always have time for a chat; the ones that always are willing to help or does it anyway; theones that always are willing to share their knowledge and experience; the ones that enjoy andorganize social activities, such as parties and beerdrinking; the ones just being good friends andcolleagues etc etc.I hope that all of those that are recognized in the above, know that they are sincerelyappreciated.

Again, I would like to thank:Mathias Uhlén, for accepting me as a PhD student and for being a very good head of thegroup. Per-Åke NygrenGGMMDD, for by all means being a very good supervisor and for letting meslide away from the protein engineering field, for critical reading of this thesis, for being a goodconference companion, etc. Joakim Lundeberg, for scientific discussions and advice, forreading of this thesis, for being a good conference companion, etc.

I also would like to thank the people at Biacore AB for good collaborations, especially BjörnPersson and Anita Larsson.

Co-authors not already mentioned are of course also acknowledged.

I also wish to express my deepest gratitude to my mother and father for all support and for thehuge amount of time spent with taking care of the girls.

Malin & Matilda, for just being there and bringing joy.

Eva, for love, endless support and understanding.

This work was supported by The Swedish Natural Science Research Council(Naturvetenskapliga forskningsrådet, NFR) and in part by Pharmacia Biosensor AB (theformer name of Biacore AB).

48

REFERENCES

Avery, O. T., MacLeod, C. M. and McCarty, M. (1944) Studies on the chemical nature of the substanceinducing transformation of pneumococcal types. Induction of transformation by a deoxyribonucleicacid fraction from Pneumococcus Type III. J Exp Med 79:137-158.

Barany, F. (1991) Genetic disease detection and DNA amplification using cloned thermostable ligase. ProcNatl Acad Sci USA 88:189-193.

Bates, P. J., Dosanjh, H. S., Kumar, S., Jenkins, T. C., Laughton, C. A. and Neidle, S. (1995) Detection andkinetic studies of triplex formation by oligodeoxynucleotides using real-time biomolecularinteraction analysis (BIA). Nucleic Acid Res 23:3627-3632.

Beck, S. and Sterk, P. (1998) Genome-scale DNA sequencing: where are we? Curr Opin Biotechnol 9:116-121.

Bellanne-Chantelot, C., Beaufils, S., Hourdel, V., Lesage, S., Morel, V., Dessinais, N., Le Gall, I., Cohen, D.and Dausset, J. (1997) Search for DNA sequence variations using a MutS-based technology. MutatRes 382:35-43.

Bianchi, N., Rutigliano, C., Tomassetti, M., Feriotto, G., Zorzato, F. and Gambari, R. (1997) Biosensortechnology and surface plasmon resonance for real-time detection of HIV-1 genomic sequencesamplified by polymerase chain reaction. Clin Diagn Virol 8:199-208.

Bier, F. F. and Scheller, F. W. (1996) Label-free observation of DNA-hybridisation and endonuclease activityon a wave guide surface using a grating coupler. Biosens Bioelectron 11:669-674.

Bier, F. F. and Fürste, J. P. (1997a) Nucleic acid based sensors. EXS 80:97-120.

Bier, F. F., Kleinjung, F. and Scheller, F. W. (1997b) Real-time measurement of nucleic-acid hybridizationusing evanescent-wave sensors: steps towards the genosensor. Sens Actuat B 38:78-82.

Björling, S., Kinjo, M., Földes-Papp, Z., Hagman, E., Thyberg, P. and Rigler, R. (1998) FluorescenceCorrelation Spectroscopy of Enzymatic DNA Polymerization. Biochemistry 37:12971-12978.

Breslauer, K. J., Frank, R., Blöcker, H. and Marky, L. A. (1986) Predicting DNA duplex stability from the basesequence. Proc Natl Acad Sci USA 83:3746-3750.

Buckle, M., Williams, R. M., Negroni, M. and Buc, H. (1996) Real time measurements of elongation by areverse transcriptase using surface plasmon resonance. Proc Natl Acad Sci USA 93:889-894.

Cai, H., Kommander, K., White, P. S. and Nolan, J. P. (1998) Flow cytometry-based hybridization andpolymorphism detection and analysis. Proc SPIE 3256:171-177.

Chee, M., Yang, R., Hubbell, E., Berno, A., Huang, X. C., Stern, D., Winkler, J., Lockhart, D. J., Morris, M. S.and Fodor, S. P. (1996) Accessing genetic information with high-density DNA arrays. Science274:610-614.

Cho, R. J., Fromont-Racine, M., Wodicka, L., Feierbach, B., Stearns, T., Legrain, P., Lockhart, D. J. andDavis, R. W. (1998a) Parallel analysis of genetic selections using whole genome oligonucleotidearrays. Proc Natl Acad Sci USA 95:3752-3757.

Cho, R. J., Campbell, M. J., Winzeler, E. A., Steinmetz, L., Conway, A., Wodicka, L., Wolfsberg, T. G.,Gabrielian, A. E., Landsman, D., Lockhart, D. J. and Davis, R. W. (1998b) A genome-widetranscriptional analysis of the mitotic cell cycle. Mol Cell 2:65-73.

Clark, L. C. and Lyons, C. (1962) Electrode system for continuous monitoring in cardiovascular surgery. AnnNY Acad Sci 102:29-45.

49

Corey, D. R. (1997) Peptide nucleic acids: expanding the scope of nucleic acid recognition. Trends Biotechnol15:224-229.

Cornell, B. A., Braach-Maksvytis, V. L., King, L. G., Osman, P. D., Raguse, B., Wieczorek, L. and Pace, R. J.(1997) A biosensor that uses ion-channel switches. Nature 387:580-583.

Cotton, R. G., Rodrigues, N. R. and Campbell, R. D. (1988) Reactivity of cytosine and thymine in single-base-pair mismatches with hydroxylamine and osmium tetroxide and its application to the study ofmutations. Proc Natl Acad Sci USA 85:4397-4401.

Cotton, R. G. H. (1993) Current methods of mutation detection. Mutat Res 285:125-144.

Cush, R., Cronin, J. M., Stewart, W. J., Maule, C. H., Molloy, J. and Goddard, N. J. (1993) The resonantmirror - a novel optical biosensor for direct sensing of biomolecular interactions. 1. Principle ofoperation and associated instrumentation. Biosens Bioelectron 8:347-353.

Danielsson, B. (1991) Calorimetric biosensors. Biochem Soc Trans 19:26-28.

Dekker, J., Kanellopoulos, P. N., van Oosterhout, J. A., Stier, G., Tucker, P. A. and van der Vliet, P. C. (1998)ATP-independent DNA unwinding by the adenovirus single-stranded DNA binding proteinrequires a flexible DNA binding loop. J Mol Biol 277:825-838.

DeRisi, J. L., Iyer, V. R. and Brown, P. O. (1997) Exploring the metabolic and genetic control of geneexpression on a genomic scale. Science 278:680-686.

Dill, K., Chan, S. D. H. and Gibbs, T. W. (1997) Detection of plasmids using DNA and RNA probes and thelight-addressable potentiometric sensor. J Biochem Biophys Methods 35:197-202.

Doyle, M. L. (1997) Characterization of binding interactions by isothermal titration calorimetry. Curr OpinBiotechnol 8:31-35.

Drmanac, R., Labat, I., Brukner, I. and Crkvenjakov, R. (1989) Sequencing of megabase plus DNA byhybridization: theory of the method. Genomics 4:114-128.

Drmanac, S., Kita, D., Labat, I., Hauser, B., Schmidt, C., Burczak, J. D. and Drmanac, R. (1998) Accuratesequencing by hybridization for DNA diagnostics and individual genomics. Nat Biotechnol 16:54-58.

Drobyshev, A., Mologina, N., Shik, V., Pobedimskaya, D., Yershov, G. and Mirzabekov, A. (1997) Sequenceanalysis by hybridization with oligonucleotide microchip: identification of beta-thalassemiamutations. Gene 188:45-52.

Ebersole, R. C., Miller, J. A., Moran, J. R. and Ward, M. D. (1990) Spontaneously formed functionally activeavidin monolayers on metal surfaces: a strategy for immobilizing biological reagents and design ofpiezoelectric biosensors. J Am Chem Soc 112:3239-3241.

Eriksson, M. and Nielsen, P. E. (1996) PNA-nucleic acid complexes. Structure, stability and dynamics. Q RevBiophys 29:369-394.

Fawcett, N. C., Evans, J. A., Chien, L.-C. and Flowers, N. (1988) Nucleic acid hybridization detected bypiezoelectric resonance. Anal Lett 21:1099-1114.

Fischer, S. G. and Lerman, L. S. (1983) DNA fragments differing by single base-pair substitutions areseparated in denaturing gradient gels: correspondence with melting theory. Proc Natl Acad SciUSA 80:1579-1583.

Fotin, A. V., Drobyshev, A. L., Proudnikov, D. Y., Perov, A. N. and Mirzabekov, A. D. (1998) Parallelthermodynamic analysis of duplexes on oligodeoxyribonucleotide microchips. Nucleic Acids Res26:1515-1521.

50

Garland, P. (1996) Optical evanescent wave methods for the study of biomolecular interactions. Q Rev Biophys29:91-117.

Gotoh, M., Hasegawa, Y., Shinohara, Y., Shimizu, M. and Tosu, M. (1995) A new approach to determine theeffect of mismatches on kinetic parameters in DNA hybridization using an optical biosensor. DNARes 2:285-293.

Gotoh, M., Hasebe, M., Ohira, T., Hasegawa, Y., Shinohara, Y., Sota, H., Nakao, J. and Tosu, M. (1997a)Rapid method for detection of point mutations using mismatch binding protein (MutS) and anoptical biosensor. Genet Anal 14:47-50.

Gotoh, M., Hasebe, M., Ohira, T. and Tosu, M. (1997b) Gene diagnosis with an affinity sensor, BIACORE--principle and applications. Rinsho Byori 45:224-228.

Griffiths, D. and Hall, G. (1993) Biosensors - what real progress is being made? Trends Biotechnol 11:122-130.

Hacia, J. G., Brody, L. C., Chee, M. S., Fodor, S. P. and Collins, F. S. (1996) Detection of heterozygousmutations in BRCA1 using high density oligonucleotide arrays and two-colour fluorescenceanalysis. Nat Genet 14:441-447.

Haruki, M., Noguchi, E., Kanaya, S. and Crouch, R. J. (1997) Kinetic and stoichiometric analysis for thebinding of Escherichia coli ribonuclease HI to RNA-DNA hybrids using surface plasmonresonance. J Biol Chem 272:22015-22022.

Hashimoto, K., Ito, K. and Ishimori, Y. (1994) Sequence-specific gene detection with a gold electrode modifiedwith DNA probes and an electrochemically active dye. Anal Chem 66:3830-3833.

Healey, B. G., Matson, R. S. and Walt, D. R. (1997) Fiberoptic DNA sensor array capable of detecting pointmutations. Anal Biochem 251:270-279.

Heller, A. (1996) Amperometric biosensors. Curr Opin Biotechnol 7:50-54.

Hutchinson, A. M. (1995) Evanescent wave biosensors. Real-time analysis of biomolecular interactions. MolBiotechnol 3:47-54.

Ito, K., Hashimoto, K. and Ishimori, Y. (1996) Quantitative analysis for solid-phase hybridization reaction andbinding reaction of DNA binder to hybrids using a quartz crystal microbalance. Anal Chim Acta327:29-35.

Jensen, K. K., Ørum, H., Nielsen, P. E. and Nordén, B. (1997) Kinetics for hybridization of peptide nucleicacids (PNA) with DNA and RNA studied with the BIAcore technique. Biochemistry 36:5072-5077.

Jin, G., Tengvall, P., Lundström, I. and Arwin, H. (1995) A biosensor concept based on imaging ellipsometryfor visualization of biomolecular interactions. Anal Biochem 232:69-72.

Jordanova, A., Kalaydjieva, L., Savov, A., Claustres, M., Schwarz, M., Estivill, X., Angelicheva, D., Haworth,A., Casals, T. and Kremensky, I. (1997) SSCP analysis: a blind sensitivity trial. Hum Mutat 10:65-70.

Jönsson, U., Fägerstam, L., Ivarsson, B., Johnsson, B., Karlsson, R., Lundh, K., Löfås, S., Persson, B., Roos,H., Rönnberg, I., Sjölander, S., Stenberg, E., Urbaniczky, C., Östlin, H. and Malmqvist, M. (1991)Real-time biospecific interaction analysis using surface plasmon resonance and a sensor chiptechnology. BioTechniques 11:620-627.

Jönsson, U. and Malmqvist, M. (1992) Real-time biospecific interaction analysis. The integration of surfaceplasmon resonance detection, general biospecific interface chemistry and microfluidics into oneanalytical system. Adv Biosensors 2:291-236.

51

Kinjo, M. and Rigler, R. (1995) Ultrasensitive hybridization analysis using fluorescence correlationspectroscopy. Nucleic Acids Res 23:1795-1799.

Kinjo, M., Nishimura, G., Koyama, T., Mets and Rigler, R. (1998) Single-molecule analysis of restriction DNAfragments using fluorescence correlation spectroscopy. Anal Biochem 260:166-172.

Kleinjung, F., Bier, F. F., Warsinke, A. and Scheller, F. W. (1997) Fibre-optic genosensor for specificdetermination of femtomolar DNA oligomers. Anal Chim Acta 350:51-58.

Kozal, M. J., Shah, N., Shen, N., Yang, R., Fucini, R., Merigan, T. C., Richman, D. D., Morris, D., Hubbell,E., Chee, M. and Gingeras, T. R. (1996) Extensive polymorphisms observed in HIV-1 clade Bprotease gene using high-density oligonucleotide arrays. Nat Med 2:753-759.

Kretschmann, E. (1971) Die Bestimmung optischer Konstanten von Metallen durch Anregung vonOberflächenplasmaschwingungen. Z Phys 241:313-324.

Landegren, U., Kaiser, R., Sanders, J. and Hood, L. (1988) A ligase-mediated gene detection technique.Science 241:1077-1080.

Lashkari, D. A., McCusker, J. H. and Davis, R. W. (1997a) Whole genome analysis: experimental access to allgenome sequenced segments through larger-scale efficient oligonucleotide synthesis and PCR.Proc Natl Acad Sci USA 94:8945-8947.

Lashkari, D. A., DeRisi, J. L., McCusker, J. H., Namath, A. F., Gentile, C., Hwang, S. Y., Brown, P. O. andDavis, R. W. (1997b) Yeast microarrays for genome wide parallel genetic and gene expressionanalysis. Proc Natl Acad Sci USA 94:13057-13062.

Liedberg, B., Nylander, C. and Lundström, I. (1983) Surface plasmon resonance for gas detection andbiosensing. Sens Actuators 4:299-304.

Liedberg, B., Lundström, I. and Stenberg, E. (1993) Principles of biosensing with an extended coupling matrixand surface plasmon resonance. Sens Actuators B 11:63-72.

Lockhart, D. J., Dong, H., Byrne, M. C., Follettie, M. T., Gallo, M. V., Chee, M. S., Mittmann, M., Wang, C.,Kobayashi, M., Horton, H. and Brown, E. L. (1996) Expression monitoring by hybridization tohigh-density oligonucleotide arrays. Nat Biotechnol 14:1675-1680.

Lorenzo, E., Pariente, F., Hernandez, L., Tobalina, F., Darder, M., Wu, Q., Maskus, M. and Abruna, H. D.(1998) Analytical strategies for amperometric biosensors based on chemically modified electrodes.Biosens Bioelectron 13:319-332.

Lukosz, W. (1995) Integrated optical chemical and direct biochemical sensors. Sens Actuat B 29:37-50.

Lundbäck, T., Hansson, H., Knapp, S., Ladenstein, R. and Härd, T. (1998) Thermodynamic characterization ofnon-sequence-specific DNA-binding by the Sso7d protein from Sulfolobus solfataricus. J Mol Biol276:775-786.

Lundström, I. (1994) Real-time biospecific interaction analysis. Biosens Bioelectron 9:725-736.

Lysov, Y. P., Florentiev, V. L., Khorlyn, A. A., Khrapko, K. R., Shick, V. V. and Mirzabekov, A. D. (1988)Determination of the nucleotide sequence of DNA using hybridization with oligonucleotides. Anew method. Dokl Akad Nauk SSSR 303:1508-1511.

Löfås, S. and Johnsson, B. (1990) A novel hydrogel matrix on gold surfaces in surface plasmon resonancesensors for fast and efficient covalent immobilization of ligands. J Chem Soc Chem Commun21:1526-1528.

Löfås, S., Malmqvist, M., Rönnberg, I., Stenberg, E., Liedberg, B. and Lundström, I. (1991) Bioanalysis withsurface plasmon resonance. Sens Actuators B 5:79-84.

52

Malmqvist, M. and Karlsson, R. (1997) Biomolecular interaction analysis: affinity biosensor technologies forfunctional analysis of proteins. Curr Opin Chem Biol 1:378-383.

McKusick, V. A. (1997) Genomics: structural and functional studies of genomes. Genomics 45:244-249.

Mikkelsen, S. R. (1996) Electrochemical biosensors for DNA sequence detection. Electroanal 8:15-19.

Millan, K. M., Saraullo, A. and Mikkelsen, S. R. (1994) Voltammetric DNA biosensor for cystic fibrosis basedon a modified carbon paste electrode. Anal Chem 66:2943-2948.

Milner, N., Mir, K. U. and Southern, E. M. (1997) Selecting effective antisense reagents on combinatorialoligonucleotide arrays. Nat Biotechnol 15:537-541.

Minunni, M., Sklàdal, P. and Mascini, M. (1994) A piezoelectric quartz crystal biosensor as a direct affinitysensor. Anal Lett 27:1475-1487.

Moser, I., Schalkhammer, T., Pittner, F. and Urban, G. (1997) Surface techniques for an electrochemical DNAbiosensor. Biosens Bioelectron 12:729-737.

Mullis, K., Faloona, F., Scharf, S., Saiki, R., Horn, G. and Erlich, H. (1986) Specific enzymatic amplificationof DNA in vitro: the polymerase chain reaction. Cold Spring Harb Symp Quant Biol 51:263-273.

Mullis, K. B. and Faloona, F. A. (1987) Specific synthesis of DNA in vitro via a polymerase-catalyzed chainreaction. Methods Enzymol 155:335-350.

Myers, R. M., Larin, Z. and Maniatis, T. (1985) Detection of single base substitutions by ribonuclease cleavageat mismatches in RNA:DNA duplexes. Science 230:1242-1246.

Napier, M. E., Loomis, C. R., Sistare, M. F., Kim, J., Eckhardt, A. E. and Thorp, H. H. (1997) Probingbiomolecule recognition with electron transfer: electrochemical sensors for DNA hybridization.Bioconjug Chem 8:906-913.

Newton, C. R., Graham, A., Heptinstall, L. E., Powell, S. J., Summers, C., Kalsheker, N., Smith, J. C. andMarkham, A. F. (1989) Analysis of any point mutation in DNA. The amplification refractorymutation system (ARMS). Nucleic Acids Res 17:2503-16.

Nickerson, D. A., Kaiser, R., Lappin, S., Stewart, J., Hood, L. and Landegren, U. (1990) Automated DNAdiagnostics using an ELISA-based oligonucleotide ligation assay. Proc Natl Acad Sci USA87:8923-8927.

Nicolini, C., Erokhin, V., Facci, P., Guerzoni, S., Ross, A. and Paschkevitsch, P. (1997) Quartz balance DNAsensor. Biosens Bioelectron 12:613-618.

Nilsson, P., Persson, B., Uhlén, M. and Nygren, P-Å., (1997). Analysis of DNA sequence using biosensors, inLaboratory methods for the detection of mutations and polymorphisms in DNA, Taylor, G. R. (ed),CRC Press Inc, Boca Raton, FL, USA. pp. 253-261.

Nolan, J. P., Shen, B., Park, M. S. and Sklar, L. A. (1996) Kinetic analysis of human flap endonuclease-1 byflow cytometry. Biochemistry 35:11668-11676.

Nolan, J. P. and Sklar, L. A. (1998) The emergence of flow cytometry for sensitive, real-time measurements ofmolecular interactions. Nat Biotechnol 16:633-638.

Okahata, Y., Kawase, M., Niikura, K., Ohtake, F., Furusawa, H. and Ebara, Y. (1998) Kinetic Measurementsof DNA Hybridization on an Oligonucleotide-Immobilized 27-MHz Quartz Crystal Microbalance.Anal Chem 70:1288-1296.

Okayama, H., Curiel, D. T., Brantly, M. L., Holmes, M. D. and Crystal, R. G. (1989) Rapid, nonradioactivedetection of mutations in the human genome by allele-specific amplification. J Lab Clin Med114:105-113.

53

Olson, J. D., Panfili, P. R., Zuk, R. F. and Sheldon, E. L. (1991) Quantitation of DNA hybridization in a siliconsensor-based system: application to PCR. Mol Cell Probes 5:351-358.

Orita, M., Iwahana, H., Kanazawa, H., Hayashi, K. and Sekiya, T. (1989) Detection of polymorphisms ofhuman DNA by gel electrophoresis as single-strand conformation polymorphisms. Proc Natl AcadSci USA 86:2766-2770.

Panayotou, G. (1998a) Surface plasmon resonance. Measuring protein interactions in real time. Methods MolBiol 88:1-10.

Panayotou, G., Brown, T., Barlow, T., Pearl, L. H. and Savva, R. (1998b) Direct measurement of the substratepreference of uracil-DNA glycosylase. J Biol Chem 273:45-50.

Pastinen, T., Kurg, A., Metspalu, A., Peltonen, L. and Syvänen, A. C. (1997) Minisequencing: a specific toolfor DNA analysis and diagnostics on oligonucleotide arrays. Genome Res 7:606-614.

Pease, A. C., Solas, D., Sullivan, E. J., Cronin, M. T., Holmes, C. P. and Fodor, P. A. (1994) Light-generatedoligonucleotide arrays for rapid DNA sequence analysis. Proc Natl Acad Sci USA 91:5022-5026.

Piunno, P. A. E., Krull, U. J., Hudson, R. H. E., Damha, M. J. and Cohen, H. (1995) Fiber-optic DNA sensorfor fluorometric nucleic acid determination. Anal Chem 67:2635-2643.

Plum, G. E. and Breslauer, K. J. (1995) Calorimetry of proteins and nucleic acids. Curr Opin Struct Biol5:682-690.

Pockrand, I., Swalen, J. D., Gordon, J. G. and Philpott, M. R. (1978) Surface plasmon spectroscopy of organicmonolayer assemblies. Surface Sci 74:237-244.

Polzius, R., Dießel, E., Bier, F. F. and Bilitewski, U. (1997) Real-time observation of affinity reactions usinggrating couplers: determination of the detection limit and calculation of kinetic rate constants.Anal Biochem 248:269-276.

Proudnikov, D., Timofeev, E. and Mirzabekov, A. (1998) Immobilization of DNA in polyacrylamide gel for themanufacture of DNA and DNA-oligonucleotide microchips. Anal Biochem 259:34-41.

Ramsden, J. J. (1997) Optical biosensors. J Mol Recognit 10:109-120.

Ratilainen, T., Holmen, A., Tuite, E., Haaima, G., Christensen, L., Nielsen, P. E. and Norden, B. (1998)Hybridization of peptide nucleic acid. Biochemistry 37:12331-12342.

Ronaghi, M., Uhlén, M. and Nyrén, P. (1998) Real-time pyrophosphate detection for DNA sequencing.Science 281:363-365.

Rutigliano, C., Bianchi, N., Tomassetti, M., Pippo, L., Mischiati, C., Feriotto, G. and Gambari, R. (1998)Surface plasmon resonance for real-time monitoring of molecular interactions between a triplehelix forming oligonucleotide and the Sp1 binding sites of human Ha-ras promoter: Effects of theDNA-binding drug chromomycin. Int J Oncol 12:337-343.

Saiki, R. K., Scharf, S., Faloona, F., Mullis, K. B., Horn, G. T., Erlich, H. A. and Arnheim, N. (1985)Enzymatic amplification of beta-globin genomic sequences and restriction site analysis fordiagnosis of sickle cell anemia. Science 230:1350-1354.

Saiki, R. K., Bugawan, T. L., Horn, G. T., Mullis, K. B. and Erlich, H. A. (1986) Analysis of enzymaticallyamplified beta-globin and HLA-DQ alpha DNA with allele-specific oligonucleotide probes. Nature324:163-166.

Saiki, R. K., Gelfand, D. H., Stoffel, S., Scharf, S. J., Higuchi, R., Horn, G. T., Mullis, K. B. and Erlich, H. A.(1988) Primer-directed enzymatic amplification of DNA with a thermostable DNA polymerase.Science 239:487-491.

54

Saiki, R. K., Walsh, P. S., Levenson, C. H. and Erlich, H. A. (1989) Genetic analysis of amplified DNA withimmobilized sequence-specific oligonucleotide probes. Proc Natl Acad Sci USA 86:6230-6234.

Sanger, F., Nicklen, S. and Coulson, A. R. (1977) DNA sequencing with chain-terminating inhibitors. ProcNatl Acad Sci USA 74:5463-5467.

Sauerbrey, G. (1959) Verwendung von schwingquarzen zur wägung dunner schichten und zur mikrowägung.Z Phys 155:206-222.

Schena, M., Heller, R. A., Theriault, T. P., Konrad, K., Lachenmeier, E. and Davis, R. W. (1998) Microarrays:biotechnology's discovery platform for functional genomics. Trends Biotechnol 16:301-306.

Sjölander, S. and Urbaniczky, C. (1991) Integrated fluid handling system for biomolecular interaction analysis.Anal Chem 63:2338-2345.

Smith, J. and Modrich, P. (1996) Mutation detection with MutH, MutL, and MutS mismatch repair proteins.Proc Natl Acad Sci USA 93:4374-4379.

Sommer, S. S., Cassady, J. D., Sobell, J. L. and Bottema, C. D. (1989) A novel method for detecting pointmutations or polymorphisms and its application to population screening for carriers ofphenylketonuria. Mayo Clin Proc 64:1361-1372.

Southern, E. M. (1975) Detection of specific sequences among DNA fragments separated by gelelectrophoresis. J Mol Biol 98:503-517.

Southern, E. M., Maskos, U. and Elder, J. K. (1992) Analyzing and comparing nucleic acid sequences byhybridization to arrays of oligonucleotides: evaluation using experimental models. Genomics13:1008-1017.

Stenberg, E., Persson, B., Roos, H. and Urbaniczky, C. (1991) Quantitative determination of surfaceconcentration of protein with surface plasmon resonance using radiolabeled proteins. J ColloidInterface Sci 143:513-526.

Stimpson, D. I., Hoijer, J. V., Hsieh, W. T., Jou, C., Gordon, J., Theriault, T., Gamble, R. and Baldeschwieler,J. D. (1995) Real-time detection of DNA hybridization and melting on oligonucleotide arrays byusing optical wave guides. Proc Natl Acad Sci USA 92:6379-6383.

Storri, S., Santoni, T., Minunni, M. and Mascini, M. (1998) Surface modifications for the development ofpiezoimmunosensors. Biosens Bioelectron 13:347-357.

Su, H. and Thompson, M. (1995) Kinetics of interfacial nucleic acid hybridization studied by acoustic networkanalysis. Biosens Bioelectron 10:329-340.

Syvänen, A.-C., Aalto-Setälä, K., Harju, L., Kontula, K. and Söderlund, H. (1990) A primer-guided nucleotideincorporation assay in the genotyping of apolipoprotein E. Genomics 8:684-692.

Szabo, A., Stolz, L. and Granzow, R. (1995) Surface plasmon resonance and its use in biomolecular interactionanalysis (BIA). Curr Opin Struct Biol 5:699-705.

Thorp, H. H. (1998) Cutting out the middleman: DNA biosensors based on electrochemical oxidation. TrendsBiotechnol 16:117-121.

Tiefenthaler, K. (1992) Integrated optical couplers as chemical waveguide sensors. Adv Biosensors 2:261-289.

Turbadar, R. (1959) Complete adsorption of light by thin metal films. Proc Phys Soc 73:40-44.

Venter, J. C., Adams, M. D., Sutton, G. G., Kerlavage, A. R., Smith, H. O. and Hunkapiller, M. (1998)Shotgun sequencing of the human genome. Science 280:1540-1542.

55

Wallace, R. B., Johnson, M. J., Hirose, T., Miyake, T., Kawashima, E. H. and Itakura, K. (1981) The use ofsynthetic oligonucleotides as hybridization probes. II. Hybridization of oligonucleotides of mixedsequence to rabbit beta- globin DNA. Nucleic Acids Res 9:879-894.

Wang, D. G., Fan, J. B., Siao, C. J., Berno, A., Young, P., Sapolsky, R., Ghandour, G., Perkins, N.,Winchester, E., Spencer, J., Kruglyak, L., Stein, L., Hsie, L., Topaloglou, T., Hubbell, E.,Robinson, E., Mittmann, M., Morris, M. S., Shen, N., Kilburn, D., Rioux, J., Nusbaum, C., Rozen,S., Hudson, T. J., Lipshutz, R., Chee, M. and Lander, E. S. (1998) Large-scale identification,mapping, and genotyping of single- nucleotide polymorphisms in the human genome. Science280:1077-1082.

Wang, J., Cai, X., Rivas, G., Shiraishi, H., Farias, P. A. M. and Dontha, N. (1996) DNA electrochemicalbiosensor for the detection of DNA sequences related to the human immunodeficiency virus. AnalChem 68:2629-2634.

Wang, J., Nielsen, P. E., Jiang, M., Cai, X., Fernandes, J. R., Grant, D. H., Ozsoz, M., Beglieter, A. andMowat, M. (1997a) Mismatch-sensitive hybridization detection by peptide nucleic acidsimmobilized on a quartz crystal microbalance. Anal Chem 69:5200-5202.

Wang, J., Rivas, G., Cai, X., Chicharro, M., Parrado, C., Dontha, N., Begleiter, A., Mowat, M., Palecek, E.and Nielsen, P. E. (1997b) Detection of point mutation in the p53 gene using a peptide nucleic acidbiosensor. Anal Chim Acta 344:111-118.

Ward, M. D. and Buttry, D. A. (1990) In situ interfacial mass detection with piezoelectric transducers. Science249:1000-1007.

Watson, J. D. and Crick, F. H. C. (1953) Molecular structure of nucleic acids. A structure for deoxyribosenucleic acid. Nature 171:737-738.

Watts, H. J., Yeung, D. and Parkes, H. (1995) Real-time detection and quantification of DNA hybridization byan optical biosensor. Anal Chem 67:4283-4289.

Welford, K. (1991) Surface plasmon-polaritons and their uses. Opt Quant Electron 23:1-27.

White, M. B., Carvalho, M., Derse, D., O'Brien, S. J. and Dean, M. (1992) Detecting single base substitutionsas heteroduplex polymorphisms. Genomics 12:301-306.

Wilchek, M. and Bayer, E. A. (1988) The avidin-biotin complex in bioanalytical applications. Anal Biochem171:1-32.

Winzeler, E. A., Richards, D. R., Conway, A. R., Goldstein, A. L., Kalman, S., McCullough, M. J., McCusker,J. H., Stevens, D. A., Wodicka, L., Lockhart, D. J. and Davis, R. W. (1998) Direct allelic variationscanning of the yeast genome. Science 281:1194-1197.

Wodicka, L., Dong, H., Mittmann, M., Ho, M. H. and Lockhart, D. J. (1997) Genome-wide expressionmonitoring in Saccharomyces cerevisiae. Nat Biotechnol 15:1359-1367.

Wood, S. J. (1993) DNA-DNA hybridization in real-time using BIAcore. Microchem J 47:330-337.

Wu, D. Y., Ugozzoli, L., Pal, B. K. and Wallace, R. B. (1989) Allele-specific enzymatic amplification of beta-globin genomic DNA for diagnosis of sickle cell anemia. Proc Natl Acad Sci USA 86:2757-2760.

Yang, M. S., McGovern, M. E. and Thompson, M. (1997) Genosensor technology and the detection ofinterfacial nucleic acid chemistry. Anal Chim Acta 346:259-275.

Yershov, G., Barsky, V., Belgovskiy, A., Kirillov, E., Kreindlin, E., Ivanov, I., Parinov, S., Guschin, D.,Drobishev, A., Dubiley, S. and Mirzabekov, A. (1996) DNA analysis and diagnostics onoligonucleotide microchips. Proc Natl Acad Sci USA 93:4913-4918.

56

Youil, R., Kemper, B. W. and Cotton, R. G. (1995) Screening for mutations by enzyme mismatch cleavagewith T4 endonuclease VII. Proc Natl Acad Sci USA 92:87-91.

Zhai, J. H., Cui, H. and Yang, R. F. (1997) DNA based biosensors. Biotechnol Adv 15:43-58.