combinatorial peptide library protocols

Introduction The four general methods to generate and screen a huge combinatorial pep- tide library +-lo7 peptides are: biological libraries such as filamentous phage I, plasmid 2 or po

Synthesis of a One-Bead One-Compound

Kit S Lam and Michal Lebl

1 Introduction

The four general methods to generate and screen a huge combinatorial pep- tide library +-lo7 peptides) are: biological libraries such as filamentous phage (I), plasmid (2)) or polysome (3) libraries; the “one-bead one-compound” syn- thetic combmatonal library method or the “Selectlde process” (4-6); synthetic peptide library methods that require deconvolution, such as an iterative approach (7,8), positional scanning (9); orthogonal partition approach (JO), or recurse deconvolution (II); and synthetic library using affinity column selection method (12,13)

There are advantages and disadvantages m each of these methods In gen- eral, the main advantages of the biological library method are that large pep- tides can be displayed on a filamentous phage library, and that large protein folds can be mcorporated into the library However, the main disadvantage is that biological libraries, in general, are restricted to all L-amino acids In con- trast, the remaining three methods all use synthetic libraries; therefore, o-amino acids, unnatural ammo acids, nonpeptide components, and small rigid scaffoldings can all be incorporated into these libraries

The “one-bead one-compound” library is based on the concept (4,5) that when a solid-phase split synthesis method (4,8,14) is used, each solid-phase particle (bead) displays only one peptide entity although there are approx 1013 copies of the same peptide in the same bead The resulting peptide-bead library (e.g., lo7 beads) is then screened in parallel using either “on-bead” binding assays (15) or “solution phase-releasable” assays (16) to identify peptide-beads with the desired biologic, biochemical, chemical, or physical properties The

From Methods m Molecular Biology, vol 87 Combmatonal Peptrde Library Protocols

2 Lam and Lebl

positive peptide-beads are then physically isolated for microsequencing with

an automatic protein sequencer In this chapter, detailed methods for the syn- thesis of a random “one-bead one-compound” combinatorial peptide library will be described Chapters 2 and 10 give examples of two general screening methods for such libraries

3 Technical grade solvents such as dimethylformamide (DMF) or dichloromethane (DCM) may be obtained from many different chemical suppliers HPLC-grade DMF for the coupling may be obtamed from Burdock and Jackson, Muskegon,

MI Ethanol, phenol, p-cresole, thioamsole, ethanedithiol, pyndme, and potas- sium cyanide may be obtained from many different chemical suppliers

2 Motorized rockmg platform

3 Randomization glass vessel (chromatography column 5-6 x 18 cm) fitted with a medmm glass smtered frit connected to vacuum and nitrogen via a two-way valve from below The three positions of the valve are “off,” “vacuum,” or “nitrogen.”

4 Recnculatmg water aspirator or a solvent-resistant vacuum pump with cold trap

5 Nitrogen tank

One-Bead One-Compound

3 Methods

3.1 Synthesis of a Linear Pentapeptide Library

As indicated earlier, a solid-phase split synthesis method (4,8,14) is used to generate a random peptide library The composition and final structure of the peptide library depends on the number of amino acids (one or more) used m each coupling cycle and the number of coupling cycles used The final peptide library may be linear or cyclic, or have specific secondary structures For sim- plicity, the method for the synthesis of a linear pentapeptide library with all 19 eukaryotic amino acids except cysteine is given below:

1 Swell 10 g TentaGel Resin S-NH, beads (- 0 25 mEq/g, see Notes 1 and 2) for at least 2 h m HPLC-grade DMF with gentle shaking in a silicomzed flask

2 Wash the beads twice with HPLC-grade DMF in the slllcomzed randomlzatlon vessel as follows* add 75 mL DMF from the top, gently bubble nitrogen from below through the smtered glass for 2 min, then remove the DMF by vacuum from below (see Note 3)

3 Transfer all the beads to a slllcomzed flask in HPLC-grade DMF Then dlstrlbute the beads into 19 equal allquots A disposable polyethylene transfer plpet IS extremely useful m the even distribution of the beads mto each polypropylene vial (see Note 4)

4 Allow the beads to settle and remove most of the DMF above the settled bead surface from each polypropylene reaction vial

5 Add threefold molar excess of each of the 19 Fmoc-protected ammo acids (see Note 5) and threefold molar excess of HOBt to each reaction vial using a mml-

6 Add threefold molar excess each of BOP and DIEA to each reaction vial to ml-

tlate the coupling reaction

7 Cap the reaction vials tightly and rock them gently for 1 h at room temperature

8 To confirm the completion of couplmg reaction, plpet a minute amount of resin

from each reaction vial into small borosilicate glass tubes (6 x 50-mm) and per- form ninhydrm test (17) as follows:

Wash the minute quantity of resin m the small glass tubes (6 x 50-mm)

(2-methylbutan-2-ol), acetic acid, t-amyl alcohol, DMF, and ether Add to each tube one drop of each of the following three reagents, (ninhydrin m etha- nol (0.1 g/mL), phenol m ethanol (4 g/mL), and potassium cyanide stock solution diluted 50 times with pyridme Place the tubes m a heating block at 120°C for 2 min Observe the color intensity of the beads under a microscope

To ensure complete couplmg, every bead from the minute quantity of sample

beads should be nmhydrin negative, I e , straw yellow color

4 Lam and Lebl

9 If the couplmg IS mcomplete (some beads remamed purple or brown with nmhy- drm test), remove the supernatant from those reaction vials and add fresh Fmoc- protected ammo acids, BOP, DIEA, and HOBt mto the reaction vial for addmonal coupling

10 If the couplmg 1s complete (beads remained straw yellow color with nmhydrm test) discard the supernatants of each reaction vial, and transfer and wash all the beads to the randomtzation vessel with technical grade DMF

11 After all the 19 couplmg reactrons are completed, all the beads are transferred to the randomizatton vessel Wash the beads (8 times, 2 mm each) with technical grade DMF

12 Add 75 mL 50% ptpertdme (m DMF) to the randomtzatton vessel to remove the Fmoc protectmg group After 10 mm, remove the ptpertdme and add 75 mL fresh 50% prpertdme After another 10 mm, wash the beads 8 times wtth techmcal grade DMF and twtce with HPLC-grade DMF

13 Distribute the beads mto each of the 19 reaction vials and carry out the next couplmg reaction as described above

14 After all the randomtzatton steps are completed, remove the Fmoc protectmg group with prpertdine as described above

15 After thorough washing with technical grade DMF (5X) followed by DCM (3X), add

10 mL of reagent K (18) to the randomrzatron vessel for 3 h at room temperature

16 Wash the deprotected resms thoroughly with DCM (3X), followed by technical grade DMF (5X), then once with 10% DIEA to neutralize the resin

17 After thorough washing with technical grade DMF, store the bead library m HPLC-grade DMF at 4°C Alternatively, the bead library can be washed thor- oughly with water and stored in 0.1 M HCl or 0.1 Mphosphate buffer with 0 05% sodmm azide

3.2 Synthesis of a Cyclic Peptide Library

The synthesis of a cyclic peptide library (disulfide bond formation) is essen- tially the same as that of the linear library except that Fmoc-Cys (Trt) is added

at the carboxyl as well as amino terminus of the linear random peptide After deprotectton, add a mixture of DMSO/Anisole/TFA (see Subheading 2.1., item 10) into the resin; incubate overnight at room temperature After thor- ough washing, store the library at 4°C as described above

4 Notes

1 We have tested several commerctally available resins for our library synthesis The two satisfactory resins are TentaGel (polyethylene grafted polystyrene beads) and Pepsyn gel (polydimethylacrylamtde beads) Overall, the TentaGel 1s prefer- able as it is nonsticky and mechanically more stable However, unlike Pepsyn gel, the level of substrtutron of each TentaGel bead is far from uniform Wtth the advent of combmatorral chemistry, we anticipate newer resins entering the market

m the near future

One-Bead One-Compound 5

2 TentaGel already has a long polyethylene linker and we do not routmely add additional linker for our library synthesis In contrast, a linker (preferably a hydrophilic lmker) is necessary for the synthesis of a peptide library with polydimethylacrylamide beads We have used Fmoc-P-alanme and/or Fmoc- aminocaprorc acid as linkers in the past However, aminocaproic acid is rather hydrophobic A polyethyleneglycol-based amino acid (Shearwater, Polymers, Huntsville, AL) is probably preferable

3 All glass vessels should be sdiconized thoroughly prior to use Besides using nitrogen bubbling through the randomization vessel to mix and wash the beads,

we have also prepared libraries in hourglass reaction vessels (Peptides Interna- tional, Louisville, KY), usmg rocking motion to mix the resins

4 Each polypropylene reaction vial should be engraved with a letter correspondmg

to a specific amino acid to ensure no mix-up during the synthesis

5 We often omit cysteines from the synthesis of linear peptide libraries to avoid the complication of intracham and/or interchain crosslinking

m Escherichia cob Biotechnology 11,1138-l 143

3 Kawasaki, G (199 1) Cell-free synthesis and isolation of novel genes and polypep- tides PCT International Patent Application W09 l/05058

4 Lam, K S., Salmon, S E , Hersh, E M., Hruby, V J , Kazmierski, W M , and Knapp, R J (1991) One-bead, one-peptide: a new type of synthetic peptide library for identifymg bgand-bmdmg activity Nature 354,82-84

5 Lebl, M., Krchnak, V., Sepetov, N F., Seligmann, B., Strop, P., Felder, S and Lam, K S (1995) One-bead-one structure combmatorial libraries Bzopolymers 37,177-198

6 Lam, K S , Lebl, M , and Krchnak, V (1997) The “one-bead-one-compound” combinatorial library method Chem Rev 97,41 l-448

7 Geysen, H M , Rodda, S J., and Mason, T J (1986) A prior-z delmeation of a peptide which mimics a discontmuous antigenic determinant Mol Immunol 23, 709-715

8 Houghten, R A , et al (199 1) Generation and use of synthetic peptide combmato- rial libraries for basic research and drug discovery Nature 354,84-86

9 Dooley, C T and Houghten, R A (1993) The use of positional scanning syn- thetic peptide combinatorial libraries for the rapid determination of opioid recep- tor ligands Life Scl 56, 1509-1517

6 Lam and Lebl

10 Deprez, B , Willard, X , Bourel, L , Coste, H , Hyafil, F., and Tartar, A (1995) Orthogonal combmatorial chemical libraries J Am Chem Sot 117,5405-5408

11 Erb, E , Janda, K., and Brenner, S (1994) Recenstve deconvolutton of combma- torial chemtcal ltbraries Proc Nutl Acad Scz USA 91, 11,422-l 1,425

12 Zuckermann, R N , Kerr, J M , Slam, M A , Banvtlle, S C., and Santa, D V (1992) Identification of highest-affinity ligands by affinity selection from eqmmo- lar pepttde mixtures generated by robotic synthesis Proc Natl Acad Scz USA 89,4505-4509

13 Songyang, Z., Carraway, K L , Eck, M J , Harrtson, S C., Feldman, R A , Mohammadi, M , Schlessmger, J , Hubbard, S R , Smith, D P , Eng, C., Lorenzo,

M J., Ponder, B A J , Mayer, B J , and Cantley, L C (1995) Catalytic spectfrc- tty of protein-tyrosme kmases 1s crmcal for selecttve stgnallmg Nature 373, 536-539

14 Furka, A., Sebestyen, F., Asgedom, M., and Dtbo, G (1991) General method for rapid synthesis of multicomponent peptide mixtures Int J Peptzde Protein Res 37,487+93

Methods f&372-380

16 Lebl, M , Krchnak, V , Salmon, S E., and Lam, K S (1994) Screenmg of com- pletely random one-bead-one-pepttde libraries for activmes m solution MethodA f&381-387

17 Kaiser, E., Colescott, R L , Bossmger, C D., and Cook, P I (1970) Color test for detection of free terminal ammo groups m the solid-phase synthesis of pepttdes Anal Blochem 34,595-602

18 King, D S., Fields, C G , and Fields, G B (1990) A cleavage method which mmtmtzes side reactions followmg Fmoc solid phase pepttde synthesis Znt J Peptlde Protein Res 36,255-266

Enzyme-Linked Calorimetric Screening

Kit S Lam

1 Introduction

In the “one-bead one-compound” combinatorial library method, each bead displays only one chemical compound although there are approx 1013 copies of the same compound in and on the same bead (I-3) With an appropriate detec- tion scheme, compound-beads with specific biological, physical, or chemical properties can be identified, and physically isolated, and then their chemical structure can be determined In biological systems, one important property that

is of interest is the binding property between a ligand and a ligate The hgate or acceptor molecule could be an enzyme (4-6)) an antibody (1,7,8), a receptor (9,10), a structural protein, or even small molecules (II) Furthermore, the

“one-bead one-compound” library method can also be applied to the discovery

of ligands that bind to the whole viral particle, bacteria, or mammalian cell by screening for compound-beads that bind to intact cells

When we mix a ligate with an “one-bead one-compound library,” some com- pound-beads may be coated by the ligate This interaction can be detected by either a labeled ligate or a labeled secondary probe that recognizes the ligate Common labels are enzyme, fluorescent probe, color dye, or radionuclide There are advantages and disadvantages to each of these methods The choice

of detection scheme depends largely on the nature and availability of specific labeled ligates From our experience, enzyme-linked calorimetric assay is prob- ably the most convenient, economical, and rapid screening method that does not require any elaborate equipment (12) Methods for the preparation of the peptide-bead library are detailed in Chapter 7 of this volume Details on the enzyme-linked calorimetric screening method will be given in the next sections,

From Methods m Molecular Bology, vol 87 Combmatonal Pep/de Library Protocols

Edlted by S CablIly 0 Humana Press Inc , Totowa, NJ

7

8 Lam

2 Materials

cal companies The following buffers are needed for the screening:

7.2, with 0.1% Tween-20 (v/v)

2 Binding Buffer 16 m&Z Na2HP04, 3 mM KH,PO,, 274 mM NaCl, 5 4 mM KCl,

pH 7 2, with 0.1% Tween-20 (v/v) and 0.1% gelatm (w/v)

3 TBS 2 5 n&Z Trts-HCl, 13 7 mM NaCl, and 0 27 mM KCl, pH 8 0

phosphate (BCIP) m 10 mL of 0 lMTris-HCl, 0 lMNaC1 with 2.34 mMMgCl,,

pH 8.5-9 0

5 Gelatin 0 1% in water

6 6M Guamdme HCl, pH 1 0

3 Methods

3.1 Screening with an Enzyme-Linked Ligate

Common enzymes used in ELISA are alkaline phosphatase, horseradish per-

oxrdase, P-galactosrdase, and glucose oxidase From our experience the alka- line phosphatase system is more specific and tends to produce the least artifact when we screen a “one-bead one-compound” library

1 If ligate-alkaline phosphatase complex is not commercially available, one may conmgate the ligate to alkaline phosphatase using bifunctional crosslmkmg reagents Many such reagents are commercially available (e g , Pierce Chemical, Rockford, IL) and standard coupling procedures are supplied by the manufactur- ers Before screening a library, one has to make sure that the coqugation method does not impair the bmdmg property of the ligate This can usually be accom- plished by an ELISA assay using a 96-well plate coated with a known hgand (see

24 h at room temperature (see Note 3)

3 Transfer the bead-library to the column and wash the beads thoroughly with PBS- Tween Then wash the bead-library one last time with TBS

Enzyme-Linked Calorimetric Library Screening 9

Fig 1 (A) Photomicrograph of a typical enzyme-linked calorimetric bead-library screen; a positive bead is noted in the middle of the micrograph (B) Single positive beads can easily be retrieved with a handheld micropipet under a dissecting microscope

4 Transfer and wash the bead-library to lo-20 polystyrene Petri dishes (100 x 20 mm) with the BCIP/alkaline phosphatase buffer (see Notes 4 and 5) More dishes may be needed if the beads are too crowded and there are too many positive beads Let the enzyme-linked color reaction develop for 30 min to 2 h Stop the reaction by acidifying the BCIP/alkaline phosphatase buffer with several drops

of 1 M HCl Figure 1A shows the photomicrograph of a typical bead-library screen

5 With the aid of a light box and a micropipet (e.g., Pipetman PlO, Gilson), transfer the turquoise beads into a small Petri dish Many colorless beads will also be transferred during this process

6 Place the small Petri dish of positive beads under a dissecting microscope and pipet individual turquoise beads to a small Petri dish of 6 M guanidine-HCI, pH

1 O (Fig 1B) At this stage, transfer only the positive beads (see Note 6) After

10 Lam

20-30 mm at room temperature m 6 Mguanidme-HCI, transfer the posmve beads

to a dish of double-dtstrlled water Then prpet each posrttve bead onto a glass filter and msert mto the protein sequencer cartrtdge for mtcrosequencmg (see

Notes 7 and 8)

3.2 Screening with an Unlabeled Ligate by Probing

with an Enzyme-Linked Secondary Antibody

1 Prepare the library as m Subheading 3.1., item 2

2 Add the alkaline phosphatase-linked anti-ligate antibody to the bead library and incubate m bmdmg buffer for l-2 h at room temperature

3 Wash the bead-library thoroughly with PBS-Tween and finally once with TBS

4 Add BCIP substrate to the library as described m Subheading 3.1., item 4

5 After 30 mm to 2 h, stop the colortmetrrc reaction by adding several drops of 1 M HCl to each Petri dish Remove all the color beads from the library over a light box with a mrcropipet These color beads interact with the secondary antibody alone and may be discarded

6 Recycle the remaining colorless library with the following steps: Incubate the library with 6 M guamdine-HCl, pH 1 O, 20-30 min, wash 5 times with double- distilled water, mix the library with DMF for 1 h, wash 5 times wrth double-distilled water, followed by PBS-Tween

7 Add the unlabeled ligate to the bead-library and incubate l-24 h at room temperature

8 Wash the bead-library thoroughly with PBS-Tween

9 Add the alkaline phosphatase-linked antrligate antibody to the bead-library and incubate l-2 h (see Note 3) Then wash the bead-library thoroughly wtth PBS- Tween and finally once with TBS

10 Add BCIP substrate to the library as described m Subheading 3.1., item 4 After

30 min to 2 h, stop the colorimetrtc reaction by adding several drops of 1 M HCl

to each Petri dish Since the library has been prescreened with the secondary antibody alone, the posrttve beads Identified at this time should be a result of bmdmg to the ligate and not to the secondary antibody

11 Isolate those individual posrtive beads for microsequencmg as descrrbed m

3 In order to mmimtze the background and false posmves, the concentration of

screening should be as dilute as possible Sometrmes lt is advantageous to use a

Enzyme-Linked Colorimetnc Library Screemng II small sample of resm (e g., 0.1 mL) to test several levels of reagent concentration before screening a large library It is not uncommon that the concentration of a reagent can be lo-fold more dilute than the optimal concentration recommended for standard ELISA

4 Although a combmatlon of BCIP and mtroblue tetrazohum (NBT) 1s commonly used m Western blot, we prefer to use BCIP alone The BCIP/NBT substrate 1s much more sensitive However, NBT can be reduced to formazon and form a dark purple preclpltate on the bead if there IS a trace amount of residual reducmg agent left in the bead-library Additionally, certain ammo acid sequences such as Asn-Asn-Asn can reduce NBT to formazon m the absence of alkaline phos- phatase Furthermore, the formazon deposit on the surface of the bead 1s msoluble

in many of the common solvents that we have tested Therefore, If BCIP/NBT substrates are used, we will not be able to recycle the library for subsequent use

or recycle a specific positive bead for confirmatory testmg before sequencing However, under certain circumstances, the tetrazohum salts are useful as a substrate as different tetrazolmm salt generates different colors upon reduction Therefore, a multicolor detection system can be designed for such applxatlons (13) Neither the formazon (when BCIPlNBT are used) nor the indigo (when BCIP alone 1s used) products ~111 affect the microsequencmg results

5 Alkalme phosphatase works best under alkaline condltlons (e g , pH 9 5) How- ever, there 1s a concern about the stability of the llgand-ligate interaction under such condltlon Therefore, depending on the ligate, we routmely adJust the BCIP/ alkahne phosphatase buffer to pH 8.5 to 9.0

6 In some instances, a dual-color colorlmetrlc detection scheme may be helpful m selectmg the true posltlve beads (13)

7 Since the rate-llmltmg step of the “one-bead one-compound” hbrary method IS the mlcrosequencmg step, one needs to ensure that most of the positive beads submltted to mlcrosequencmg are “true positives.”

8 To further improve the probability of true positlvlty, one may decolorize the posl- tive beads with DMF and restam the beads m the presence or absence of a com- peting llgand

Acknowledgments

This work was partially supported by NIH grants CA23074 and CA17094

Kit S Lam is a scholar of the Leukemia Society of America

References

1 Lam, K S., Salmon, S E , Hersh, E M , Hruby, V , Kazmlerskl, W M , and

Knapp, R J (1991) A new type of synthetic peptlde hbrary for ldentlfymg hgand-

bmdmg actlvlty Nature 354,82-84

2 Lebl, M , Krchnak, V , Sepetov, N F , Seligmann, B , Strop, P , Felder, S , and

37,177-198

Lou, Q., Leftwich, M., and Lam, K S (1996) Identification of GIYWHHY as a novel peptide substrate for human p60c-src protein tyrosme kmase Bloorg Med Chem., 4,677-682

Lam, K S., Lebl, M., Krchnak, V , Wade, S , Abdul-Lattf, F , Ferguson, R , Cuzzocrea, C , and Wertman, K (1993) Discovery of D-ammo acid contammg ligands with Selectide Technology Gene 137,13-16

Lam, K S , Lake, D , Salmon, S E , Smtth, J., Chen, M-L., Wade, S., Abdul- Latrf, F , Leblova, Z , Ferguson, R D., Krchnak, V , Sepetov, N F., and Lebl, M (1996) A one-bead, one-pepttde combmatorial library method for B-cell epitope mapping Methods A Compamon to Methods m Enzymology 9,482-493

11 Lam, K S , Zhao, Z G., Wade, S , Krchnak, V , and Lebl, M (1994) Identtfica- tion of small peptides that interact specifically with a small organic dye Drug Dev Res 33,157-160

Methods 6,372-380

13 Lam, K S., Wade, S., Abdul-Latif, F , and Lebl, M (1995) Application of a dual color detection scheme m the screening of a random combmatorial peptide library

J Immunol Methods 180,219-223

of combinatorial libraries represent a dramatic advance in the drug discovery process by greatly reducing the time needed to identify new drug leads Posi- tional scanning (PS) SCLs (Z,2) represent a modified format of the origmal synthetic combmatorial libraries described by this laboratory (3) In contrast to the original libraries, which required several iterative syntheses to identify individual active compounds, this library format provides mformation on the substituent responsible for activity at each varied position withm a structure Therefore, only a single subsequent synthesis is required The screening of PS-SCLs, in most instances, permits the identification of the most active sub- stituents at each position of a compound in a single assay Thus PS-SCLs serve

to reduce further the time required to identify new drug leads PS-SCLs are composed of individual positional SCLs, in which a single position is defined with one substituent while the remaining positions are composed of mixtures

of substituents The defined position is “walked” through the entire sequence

of the PS-SCL Therefore, the number of positional SCLs is equal to the num- ber of residues in each compound of the PS-SCL It should also be noted that each posrtional SCL, although addressing a single positron of the sequence, represents the same collection of individual compounds For example, a hexapeptide, or a compound with six positions, can be represented as: OIXXXXX, XO,XXXX, XX03XXX, XXO,XX, XXXX05X, or XXXXX06

From Methods m Molecular Biology, vol 87 Combmatonal Peptrde Library Protocols

Edlted by S Cablily 0 Humana Press Inc , Totowa, NJ

13

14 Dooley and Houghten

Using 20 amino acids (this represents 120 mixtures in total), each peptide mix- ture contains 3.2 million (205) different sequences, the six positional libraries each contam 64 million hexamers Peptide PS-SCLs can be prepared with an acetylated N-terminus, as well as with a C-terminal amide or carboxylate One highly advantageous characterlstlc of the PS-SCLs prepared in this laboratory

is that they are free to interact m solution, i.e., they are not bound to any sup- port (beads, glass, phage, and so on), and therefore can be readily screened in any assay system When used in concert, the data derived from each positional SCL yield Information about the most important substituents for every posl- tlon The information IS then used to synthesize individual compounds repre- senting all possible combmations of the most active substituents at each position This serves to confirm the PS-SCL screening results, as well as to Identify the individual compounds with the highest activities

The preparation of a PS-SCL composed of L-ammo acid hexapeptides is described This library consists of six separate positional SCLs, each composed

of 20 different peptlde mixtures having a single posltlon defined with one of the 20 natural ammo acids (represented as 0)) and the remaining five positions are composed of mixtures of 19 amino acids (represented as X; cysteine is omitted, see Note 1) The six posltlonal SCLs differ only in the location of the defined position A description IS given for the preparation of the library using either Boc or Fmoc chemistries The choices of procedure depends on the labo- ratory facilities available, safety, and financial considerations Methods for screening such a library in a radioreceptor assay are given as an example Although the methods described here involve the use of peptides, the posi- tlonal scanning concept may be equally applied to a library of any class of compounds m which there are a number of positions that may be systemati- cally altered PS-SCLs have been prepared composed of decapeptides (4), hexapeptides comprised solely of D-ammo acids (5)) of tetrapeptides comprised

of more than 50 L-, D-, and unnatural amino acids (6) and of heterocycles (24) PS-SCLs have been used successfully to identify antigenic determinants rec- ognized by monoclonal antibodies, trypsm inhibitors, opioid receptor llgands, and antlmicroblal compounds A series of papers on the use of peptlde, peptidomlmetic, polyamme and heterocyclic PS-SCLs in the aforementioned assays have been published by our laboratory (5-15,24)

2 Materials

2.1 Library Synthesis

2.1.1 General Requrements for Synthesis

1 Resin packets (T-bags) are made with polypropylene mesh (74 pm, Spectrum, Houston TX), using an impulse sealer

Synthesis and Screenmg 15

2 T-bags are filled with polystyrene resin, 200 mg, 0.2 mEq

3 Solvents for all synthetic procedures are dimethylformamide (DMF) and/or

4 A lyophihzer and somcator are used m the final preparations of the peptide mixtures

2.1.2 t-Boc Synthesis

protecting groups: benzyl is used as the side chain protection for Asp, Glu, Ser, and Thr; 2,4-dmitrophenyl for HIS; Z,chloro-benzyloxycarbonyl (CBZ) for Lys; formyl for Trp, sulfoxide for Met, p-tosyl for Arg; and 2,bromo- benzyloxycarbonyl for Tyr

3 DCM and isopropanol (IPA) are used alternatively m wash steps

4 Trifluoroacettc acid (TFA) is used to remove Boc protecting groups

5 Dusopropylethylamme (DIEA) is used as a base m neutralization steps

coupling reagents

hydrogen fluoride (HF) gas are used for side cham deprotection, and methanol is used m the wash procedure

8 HF gas and a 24-vessel cleavage apparatus are used for peptide cleavage

2.1.3 Fmoc Synthesis

1 Polyoxyethylene-grafted polystyrene resin (TentaGel)

linker

3 This synthesis employs Fmoc protected ammo acids with the following side chain protection groups: t-butyl for Ser, Thr, Tyr, Asp, and Glu, trityl for Cys, His, Asn, and Gln, Boc for Lys, and 2,2,5,7,8-pentamethylchroman-6-sulfonyl (Pmc) for Arg

4 DMF is used in the wash steps

5, Pipendme is used to remove Fmoc protecting groups

6 DIC and HOBt are used as coupling reagents

7 TFA, trusobutylsilane, water, and “Quick Snap” plastic tubes equipped with a smtered bottom disc (Isolab, Akron, OH) are used for side chain deprotectlon/ cleavage

8 Tert-butylmethylether, hexane and a benchtop centrifuge are required to precipi- tate and collect peptides

2.2 Screening

2.2.7 Receptor Assay

1 Aqueous buffer, e g ,50 mMTris, pH 7 4

2 Receptor preparation

76 Dooley and Houghten

3 1-mL Polypropylene tubes wtth caps (Contmental Laboratory Products, San Diego, CA) and 96-well trays (Costar, Pleasanton, CA)

For general procedures on solid phase peptide synthesis, readers are referred

to (16) and (17) The peptide mixtures making up the PS-SCLs are synthesized

by simultaneous multiple peptide synthesis (SMPS) (18) Mixture posrtions (X) are incorporated by couplmg mixtures of protected ammo acids for pep- ttdes, or aldehydes, carboxylic acids, and so forth for nonpeptides, using isokmetic coupling of an excess of a predetermined molar ratio, which com- pensates for the different couplmg rates of the various amino acid derivattves (Table 1) The advantage of using SMPS (also referred to as the T-Bag tech- nology, US Patent No 4,63 1,211) is that all wash and deprotection steps may

be carried out in a common vessel For the hexapeptide library, 120 T-bags (resin packets) are made and labeled (see Note 2)

3.1.1 t-Boc Synthesis

1 All bags are washed (approx 4 ml/bag for l-m-square bags) 1X DCM, 2X IPA, 2X DCM This 1s to ensure that the bags do not leak

2 Neutralize bags* 3 x 5% DIEA/DCM, 2 mm; 2X DCM, 2X DMF, 1 mm each

3 Activatton/couplmg (see Subheading 3.1.3 for couplmg procedure and Note 3)

4 Wash bags 1X DMF, 2X DCM

5 Deprotect using a 55% TFA/DCM solution for 30 mm

6 Wash bags 1X DCM, 2X IPA, 2X DCM

7 Steps 2-6 are repeated for the required number of couplmgs

8 Deprotect the peptide srde chains using (a) DNP removal, 2 5% thtophenol/DMF for 1 h Wash bags 3X DMF, 12X alternating washes of IPA and DCM; and (b) low-HF 60% DMS, 5% EDT, 10% p-cresol, and 25% HF for 2 h at 0” C Wash bags, 8X alternating washes of IPA and DCM, 4X DMF, 3X DCM, 1X methanol (MeOH)

9 Cleave peptrdes from the resm using 7 5% amsole/HF for 1 h at 0” C

10 Extract pepttdes with water or drlute acetic acid Lyophthze peptide solutions twice and reconstitute in water at lo-20 mg/mL (see Note 4) Mixtures may be stored for prolonged pertods at -20” C (see Note 5)

3.1.2 Fmoc syn thesrs

1 Wash bags 3X DMF, 3X DCM

2 Couple TFA cleavable lmker to resin (100 mEq), shake overmght

Synthesis and Screenmg 17

Table 1

Molar Ratios of Amino Acids Used for Coupling Mixture Positions (X)

3 Wash bags 5X DMF, 1 min

5 Wash bags 5X DMF (1 mm)

6 Acttvatlon/Couplmg (see Subheading 3.1.3 for coupling procedure) Fmoc- ammo acld/DIC/HOBt (5 Eq solution m DMF), 90 mm Test for coupling comple- tion (see Note 3)

7 Wash bags 5X DMF, 1 mm

8 Repeat steps 4-7 as necessary

9 Remove resin from bags and place m “Qmck Snap” plastic tubes

10 Cleave peptides using 1 5 mL of TFA/DCM/H,O/triisobutylsilane 70:20*5.5, for

3 h at room temperature

11 Snap off the tips of the “Quick Snap” tubes and add cleavage solutton to centrt- fuge tubes Precipitate pepttdes with cold (4” C) tert-butylmethylether (30 mL) Centrifuge at 3000g for 10 mm Dissolve peptides m 15 mL of water, lyophllize pepttde soluttons twice, and reconstitute m water at lo-20 mg/mL (see Note 4)

Mtxtures may be stored for prolonged periods at -20” C (see Note 5)

18 Dooley and Houghten

Table 2

Coupling Procedure for a Hexapeptide PS-SCL

“where 0 = A, or C, or D or Y See Subheading 3.1.3

bwhere X = A, and D, and E See Subheading 3.1.3

2 The 20 bags bemg coupled as 0 (1 of the 20 ammo acids) are separated mto 20 vessels and mdividually coupled to each of the 20 ammo acids (1 = A, 2 = C,

3 = D, and so on) using 0 2 A4 of ammo actd/DCM (6 mL) (with an equimolar concentration of HOBt for asparagine and glutamme), and 0.2 M DIC/DCM (6 mL) for 1 h (6 Eq)

3 Bags being coupled as X are combined m a single vessel and coupled using a mixture of 19 amino acids in ratios described in Table 1 Solutions of ammo acid mixture, DIC, and HOBt (0.5 M, solubihzed m DMF) are mixed to yield a final concentration of 0.167 M

4 The N-terminus of the peptides may be acetylated if desired usmg 0 2 M acetyl tmidazole in DMF (20 Eq)

3.2 Screening and Analysis

3.2.1 Radioreceptor Assay

1 Screening of SCLs requires a high-throughput assay A 96-well or ELISA format

is recommended Depending on the type of receptor or radiolabel used, adaptmg

a known protocol to a 96-well format may require several experiments to opti- mize assay conditions It is important to have good separation between total bmd-

mg and nonspecific bmdmg (ZlOOO cpm) and httle variation between replicates (l-5%)

2 Prepare receptor preparation, a membrane-bound receptor m tissue homogenate

is described here as an example Protein content of crude homogenates should be determined using the methods described by Bradford (19) or Lowry (20), as appropriate (see Note 6)

Synthesis and Screenmg

3 Perform the bmdmg assays m 1 -mL polypropylene tubes A sample plan for this assay IS given m Fig 1 Usually two replicates are sufftclent for screenmg stud- ies Minimtze ptpetmg by using multichannel pipets or computertzed ptpetmg stations,

4 Determine mterassay and intra-assay variation using standard curves, incubate the radlohgand m the presence of a range of concentrations of an unlabeled hgand (Stan- dard Curve) Reserve one column of each plate for a standard curve (see Fig 1)

20 Dooley and Houghten

5 Add peptide mixture (50 pL of 5 mg/mL solution) and the appropriate volumes

of buffer, radtohgand, and receptor preparation to each tube Because of the large size of most assays, it IS recommended that the receptors be added last (see Note 7)

6 Incubate assay tubes until equtlibrmm is reached This is generally longer than the time needed for the unlabeled ligand to reach equillbrmm, and needs to be determmed before the assay (see Note 8)

7 Termmate the reaction by filtration through GF-B filters on a Tomtec Harvester (see Note 9) Wash each sample on the filter with 6 mL of Tris-HCl buffer, at 4” C Count bound radioactivity on a LKB Beta Plate Liqurd Scrntillation Counter, the counts are expressed m counts per mmute (CPM)

8 Process the raw data using spreadsheet software (lotus, Excel, and so forth) Average replicates and express as percent mhtbition* 100 - [(Mean - NSB)/ (TB - NSB) * 1001 Graph data such that there are six graphs (one for each posl- tton of the hexamer); each graph should contain 20 bars (one for each ammo acid) For certam assays, this data 1s sufficient to identify one, two, or three ammo acids from each of the SIX posittons, which can then be combined to make mdi- vidual peptides (Jee step 9 below) For receptor assays, often too many mixtures are active Calculate ICsD values (concentration of mixture that mhiblts 50% radioltgand binding) in order to determine the most active mixtures

9 Perform competitive mhibition assays as above using serial dtlutions of the pep- tide mixtures Prepare five threefold dilutions and use the peptide mixture such that six concentrations are tested (e g , 1X mixture, 0.3X mixture, 0.1X mixture,

0 037X mixture, 0.012X mixture, and 0.004X mixture) Calculate IC5a values using the curve-fittmg software, e g , GRAPHPAD (ISI, San Diego, CA) For small combmatorial libraries such as the hexapepttde PS-SCL, IC,, values can easily be determined for all 120 mixtures Rank order the IC,, values for each of the SIX positions, and use these values to choose ammo acids for individual peptides

10 Synthesize all combinations of the most active mixture(s) for each of the six positrons as individual peptides (Table 3) The numbers of individual peptides to

be synthesized rises exponentially with the number of amino acids chosen (1 e , one ammo acid from each posmon generates one peptide, two amino acids from each position generates 64 [26] peptides, and three ammo acids for all SIX posi- tions generates 729 [36] peptides)

11 It is important to note that the activity of the individual pepttdes either supports

or disproves the connectivity of the most active ammo acid at each position found from the screenmg of the library For example, one of the most active peptide mixtures from the screenmg of a PS-SCL usmg ELISA was AC-XXXPXX-NH,, although none of the resulting individual peptides containing prolme at the fourth position were found to be acttve However, completron of the iterative process for this pepttde mixture yielded an active indtvtdual pepttde that was completely different from the sequences derived from the PS-SCL (21)

Synthesis and Screenmg

Table 3

21

Example of Individual Sequences Derived from Screening

a Hexapeptide Positional Scanning Library

Posltlon Amino acids chosen Pepttde sequences

Care must be taken when analyzmg data of mixtures containmg cysteme, whtch

activity, tt 1s best to iterate (sequentially define the X posmons) or prepare a posittonal scanning hbrary of that mixture

The resm swells and shrinks depending on solvent used Bags should be made large enough to accommodate swelling

It 1s important to ensure that all couplmg reactions go to completion It is htghly recommended that ninhydrm (22) or bromophenol blue (23) monitormg be car- rted out on control resins after each coupling

Somcation is used to solubdize mixtures containing hydrophobic amino acids m the defined positions It 1s important to keep the water m the somcator cool, add ice if necessary

It 1s recommended that the library be altquoted before storage to avoid freeze and thaw damage Addmonally, if the library is allquoted mto a 96-well format, the ptpeting required for screening 1s substantially reduced

Lowry Method: To 0.1 mL of sample add 0.1 mL of 2N NaOH Boll at 100” C for

10 min Cool and add 1 mL of reagant (100 ~012% w/v Na,COa m water, 1 vol 1% w/v CuSo4*5Hz0 m water, and 1 ~012% sodium potassium tartrate m water) Incubate for 10 mm Add 0.1 mL of Folm reagant, mix, and mcubate for 30 mm Read absorbance at 750 and/or 550 nm Prepare a standard curve using bovine serum albumin (BSA) and use tt to determine the concentratron of the sample Bradford Method Dissolve 100 mg Coomassie Blue G250 in 50 mL of 95% ethanol, mix with 100 mL of 85% phosphoric acid, and make up to 1 L with water Filter the reagent Ptpet 0 1 mL of sample into test tube, add 5 mL of reagent, and mix Measure absorbance at 595 nm lo-60 min after mtxmg Prepare a standard curve using BSA and use it to determine the concentration of the sample

22 Dooley and Houghten

7 Screenmg of this hbrary m an optold radioreceptor assay was optimal at a fmal concentration of 0 4 mg/mL When screening a new receptor, it is advisable to screen at a high concentration (l-5 mg/mL) and subsequently decrease the con- centration if too many mixtures are found to be active

8 The l-mL polypropylene tubes come with plugs m strips of eight We have found that the only way to ensure adequate mtxmg of components 1s to cap all tubes and invert the tray two or three times, tapping both ends to dislodge any solvent from the top or bottom of the tube

9 We have found that soaking filters in polyethylemmme (PEI), as 1s often recom- mended to reduce nonspecific binding, causes problems when using the filters obtained from Wallac PEI causes the ink and portions of the filter to stick to the harvester A 5-mg/mL BSA/buffer solution has generally sufficed to muumlze nonspecific bmdmg

2 Dooley, C T and Houghten, R A (1993) The use of postttonal scanning syn- thetic peptide combmatorial libraries for the rapid determmation of opioid recep- tor hgands Lzfe Scz 52, 1509-15 17

3 Houghten, R A., Pnulla, C., Blondelle, S E., Appel, J R., Dooley, C T., and Cuervo, J H (199 1) Generation and use of synthetic peptide combmatorial librar- ies for basic research and drug discovery Nature 354,84-86

4 Pmllla, C., Appel, J R , and Houghten, R A (1994) Investigation of antigen- antibody mteractions using a soluble nonsupport-bound synthetic decapeptide library composed of four trillion sequences Bzochem J 301,847-853

5 Pimlla, C , Appel, J R , Blondelle, S E., Dooley, C T., Elchler, J , Ostresh, J M.,

and Houghten, R A (1994) Versatility of positional scannmg synthetic combma- torial libraries for the JdentifJcatlon of mdtvidual compounds Drug Dev Res 33, 133-145

6 Dooley, C T , Bower, A N , and Houghten, R A (1996) Identification of mu-selective tetrapeptides using a positional scannmg combmatorial library con-

Biology (Proceedmgs of the Fourteenth American Peptzde Symposium (Kaumaya,

P T P and Hodges, R S., eds ), Escom, Leaden, Germany, pp 623,624

7 Pmilla, C , Buencammo, J., Appel, J R., Houghten, R A , Brassard, J A., and Ruggeri, Z M (1995) Two antipeptide monoclonal antibodies that recogmze

adhesive sequences in fibrinogen Identification of antlgenlc determinants and

Synthesis and Screenmg 23 unrelated sequences using synthetic combmatorial libraries Blamed Pept Pro- terns Nucletc Actds 1, 199-204

8 Pimlla, C Buencamino, J Appel, J R., Hopp, T P , and Houghten, R A (1995) Mapping the detailed speciftctty of a calcmm-dependent monoclonal antibody through the use of soluble positional scanning combmatonal libraries identifica- tion of potent calcium-independent antigens Mol Diversity 1,21-28

9 Ostresh, J M., Husar, G M., Blondelle, S E., Dorner, B , Weber, P A., and Houghten, R A (1994) “Ltbrartes from libraries” Chemical transformatton of combmatorial libraries to extend the range and repertoire of chemical diversity Proc Nut1 Acad Scl USA 91,11,138-l 1,142

10 Perez-Pay&E , Takahashi, E., Mmgarro, I., Houghten, R A., and Blondelle, S E (1996) Use of synthetic combmational libraries to identify pepttde mhtbitors of Ca2’-complexed calmodulm, m Peptides: Chemistry, Structure and Btology (Pro- ceedings of the Fourteenth American Peptzde Symposium) (Kaumaya, P T P and Hodges, R S., eds ), Escom, Leiden, Germany, pp 303,304

11 Perez-Pay& E., Houghten, R A., and Blondelle, S E (1996) The destgn of self- assemblmg pepttde complexes usmg conformattonally defined ltbrartes, m Tech- niques tn Protem Chemtstry VII (Marshak, D., ed ), Academic Press, San Diego,

CA, pp 65-7 1

12 Blondelle, S E , Houghten, R A , and Perez-Pay& E (1996) All D-ammo acrd hexapepttde mhlbttors of meltttrn’s cytolyttc activity derived from synthetic com- bmatonal bbraries J Mol Recog 9, 163-168

13 Blondelle, S E.,Takahasht, E , Houghten, R A., and Perez-Pay& E (1996) Rapid rdentifrcation of compounds having enhanced antimicrobial activity using confor- mattonally defmed combmatortal ltbrartes Bzochem J 313, 141-147

14 Dooley , C T and Houghten, R A (1995) Identtficatton of mu-selecttve poly- amine antagonists from a synthetic combmatorial library, Analgesza, 1,400404

15 Eichler, J., Lucka, A W , and Houghten, R A (1994) Cyclic pepttde template combmatorial libraries Synthesis and identification of chymotrypsin inhibitors, Pept Res 7,300-307

16 Steward, J M and Young, J D (1984) Soled Phase Pepttde Syntheses 2nd ed , U.S.A Pierce Chemtcal Company, Rockford, IL

17 Pennmgton, M W and Dunn, B M., eds (1994) Pepttde Synthesis Protocols Methods in MoEeculur Biology Humana, Totowa, NJ

18 Houghten, R A (1985) General method for the raptd solid-phase synthesis of large numbers of peptides specificity of antigen-antibody mteraction at the level

of mdividual ammo actds Proc Nat/ Acad Set USA, 82,5 13 1-5 135

19 Bradford, M M (1976) A rapid and sensitive method for the quantitation of mrcrogram quanttties of protem utibzmg the princtple of protein-dye binding Anal Btochem 72,248-254

20 Lowry, 0 H , Rosebrough, N J , Farr, A L., and Randall, R J (195 I) Protein measurement with the Folm phenol reagent J Btol Chem 193,265-275

21 Pimlla, C , Buencammo, J., Houghten, R A , and Appel, J R (1995) Detailed studies of antibody specificity using synthetic combinatorial libraries, m Vuccznes

24 Dooley and Houghten

1995: Molecular Approaches to the Control of Infectlow Diseases (Brown, F , Chanock, R., Ginsberg, H , and Non-by, E., eds.), Cold Sprmg Harbor Laboratory Press, Cold Spring Harbor, pp 13-17

22 Kaiser, E T., Colescott, R L., Blossmger, C D , and Cook, P I (1970) Color test for detection of free terminal ammo groups m the solid-phase synthesis of pep- tides AnaE Blochem 34,595-598

23 Krchnak, V , Vagner, J., Safir, P,, and Lebl, M (1988) Nonmvasive contmuous momtormg of solid phase peptide synthesis by acid-base mdtcator Collect Czech Chem Commun 53,2542-2548

24 Nefzi, A., Ostresh, J M , Meyer, J -P , and Houghten, R A (1997) Solid phase synthesis of heterocycltc compounds from linear pepttdes cyclic ureas and thioureas Tetrahedron Lett 38,93 l-934

4

Synthesis and Screening of Peptide Libraries

Achim Kramer and Jens Schneider-Mergener

1 Introduction

There are different strategies for the construction of soluble and solid phase- bound chemrcal peptide libraries These libraries have been used for the detec- tion of epitopes as well as for the identification of peptides that act as antagonists of medically relevant proteins We have prepared different types of cellulose-bound peptide libraries by spot synthesis (I), which is a powerful tool for the simultaneous and positronally addressable synthesis of thousands

of peptrdes or peptide mixtures bound to continuous cellulose membrane sup- ports Presently up to 8000 different spots (peptides or peptide mixtures) can

be automatically synthesized on a 20 x 30-cm cellulose membrane and screened for ligand binding within l-2 wk

These libraries can be used for the detection of peptides that bind to pro- teins, metals, and nucleic acids (2) We have mapped several linear and nonlm- ear antibody epitopes (3-9), and also used this approach for the detection of receptor-ligand interactions For instance, cellulose-bound peptide scanning libraries allowed the detection of the contact sites between tumor necrosis fac- tor-o, and interleukin-6 with its receptors (4,IO) Furthermore, this method proved to be valuable to characterize heat shock protein-peptide mteractlons (II) As another biologically relevant application, these libraries were applied for the study of metal-peptide interactions For example, we identified tech- netium-99m binding peptides important for tumor diagnosis (12) and nickel- binding peptides that can be used for the purification of recombinant proteins (3) The synthesis of cellulose-bound peptide libraries is not restricted to L-amino acids Other building blocks, such as o-amino acids, unnatural ammo acids, and organic compounds, can also be used Furthermore, the synthesis of

From Methods m Molecular Bology, vol 87 Combmatonal Pepbde Library Protocols

Edlted by S CablIly 0 Humana Press Inc , Totowa, NJ

25

26 Kramer and Schneider-Mergener

Cellulose modification (3.1)

Amino acid coupling (3.4):

Side group

El cleavage (3.6)

Rg 1 Outhne of the synthesis procedure of cellulose-bound pepttde libraries

cyclic and branched peptrde libraries can be achieved on cellulose membranes (3,13,14)

Described here are protocols for the manual synthesis of a combmatorral peptrde library XXB,B2XX (B = defined position, X = randomized position) (2,15,16), a peptide scanning library, and a mutational analysis of a peptrde eprtope An outline of the synthesis strategy IS given m Fig 1 Furthermore,

we provide protocols for the screening of these libraries with protein ligands

As an example, we have screened the libraries with the monoclonal anti- transforming growth factor-a (TGFa) antibody Tab 2 (Fig 2) The combina-

Cellulose Membrane Supports 27

Fig 2 Reaction of the anti-TGFa antibody Tab 2 with different types of peptide libraries (A) Combinatorial library XXB,B,XX, (B) mutational analysis of the TGFa

acids overlapping starting with the upper left spot)

torial peptide library XXB ,B,XX allowed the a priori delineation of the TGFa epitope The peptide scanning library consisting of overlapping peptides span- ning the entire TGFa sequence also led to the detection of the epitope In addi- tion, a mutational analysis substituting each single epitope residue by all 20 gene-encoded amino acids gave interesting insights into the molecular nature

of this peptide-antibody recognition This technique allowed the identification

of amino acids that cannot be substituted by other residues and are therefore key residues in antibody binding

A short chapter introduces the Auto-Spot Robot APS 222 (Abimed GmbH, Langenfeld, Germany), which accelerates the spotting steps, thus allowing the semiautomated, parallel synthesis of a high number of peptides The software for generation of library sequence files, described in this paper, can be purchased from Jerini BioTools GmbH (Berlin, Germany), which also manu- factures synthesized libraries

2 Materials

The amino acids (one letter code) are 9-fluorenylmethoxycarbonyl (Fmoc)- protected With the exception of Fmoc-P-alanine-OH (Subheading 2.1.), all amino acids are activated with either pentafluorophenyl (Pfp) or 3-hydroxy- 2,3-dihydroxy-4-oxobenzotriazolyl (Dhbt) The following side-chain protect- ing groups are used: trityl for C, H, N, and Q; t-butyl for D, E, S, and T;

28 Kramer and Schneider-Mergener

t-butoxycarbonyl for K and W; and pentamethylchroman sulfonyl for R All

other chemicals should be purchased in their purest quality and used without further purification (see Note 1) All peptide synthesis steps should be carried

out in a well-ventilated hood and wtth sufficient protective clothing

2.1 Modification of the Cellulose

1 Cellulose paper: Whatman 50 (Whatman, Maidstone, England)

2 Activated p-alamne solution: 0 2 M Fmoc-P-alanme-OH activated with 0 24 A4 DIC (dnsopropylcarbodlimlde) and 0 4 M NM1 (N-methyhmldazole)

3 DMF dlmethylformamlde

4 Piperidme solution: 20% plperldme m DMF

5 Methanol

2.2 Definition of the Spots

1 Fmoc-P-alanme-OPfp solution 0 3MFmoc-P-alanm-OPfp m DMSO (dlmethyl- sulfoxide)

2 DMF

3 Plperidine solution 20% plperidine m DMF

4 Methanol

5 Acetanhydrlde solution A* 2% acetanhydrlde m DMF

6 Acetanhydrlde solution B 2% acetanhydnde, 1% DIPA (dusopropylethylamme)

3 Plperldine solution 20% plperldme m DMF

2.4 Coupling of the Amino Acids

1 Amino acid solutions* each of the 20 genetlcally encoded ammo acids 1s used as

0 3 A4 solution m NMP (N-methylpyrrolldone) These solutions are stable at -2O’C for several days with the exception of the active ester of argmme, which has to be freshly prepared each working day

2 X-mixture eqmmolar mixture of 17 ammo acids (all 20 genetically encoded ammo acids except cysteme, methlonme, and tryptophane) Use a concentration

of 1.5 x (functionality of the spot) per yL Make use of the result of the calcula- tion of Subheading 3.3 For a functlonahty of 60 nmol per spot, the concentra- tion of the X-mixture has to be 90 mM m NMP MIX the 0 3 M amino acid solutions and dilute with NMP

3 DMF

4 Piperldine solution* 20% plperldme m DMF

5 BPB-solution A* 0.01% (w/v) bromophenol blue m methanol

Cellulose Membrane Supports 29 2.5 Acetylation of the N-terminus

2 DMF

3 Methanol

2.6 Cleavage of the Side-Chain Protecting Groups

1 Deprotectlon solution: 50% TFA (tnfluoroacetic acid), 3% trusobutylsllane, 2% water, 1% phenol m DCM (dlchloromethane) TFA 1s toxic and very corrosive and should be handled with the greatest caution Do not mix TFA and DMF waste

as it can undergo an exothermlc and explosive reaction Consult your safety officer for approved handling and disposal procedures

2 DCM dlchloromethane

3 DMF

4 Methanol

2.7 Automated Spot Synthesis

1 Auto-Spot Robot APS 222 (Ablmed GmbH, Langenfeld, Germany)

2 Software DIGEN (Jerml BloTools GmbH, Berlin, Germany)

3 Peptlde synthesis chemicals of the previous sections

2.8 Synthesis of a Mutational Analysis

Peptide synthesis chemicals of the previous sections

2.9 Synthesis of a Peptide Scanning Library

Peptide synthesis chemicals of the previous sections

2 IO Screening of the Peptide Libraries

1 Methanol

137mM NaCl, 2 7 mA4 KCl Adjust the pH to 8 0 with HCl

3 Blocking buffer: dilute blockmg reagent (CRB, NorthwItch, UK) 1 *lo m T-TBS and add 10% (w/v) sucrose

5 Ligand solution: dilute the protein of interest to a final concentration of 0.1-l pg/mL m blocking buffer (see Note 3)

2.11 Detection of Ligand Binding (Alkaline Phosphatase Method)

2 Blocking buffer dilute blocking reagent (CRB) 1.10 m T-TBS and add 10% (w/v) sucrose

3 Alkaline phosphatase-labeled antlbody solution ddute an alkalme phosphatase- conjugated antibody, which 1s directed against the primary hgand, 1 lO,OOO m blockmg buffer (see Note 3)

30 Kramer and Schneder-Mergener

4 Nttroblue tetrazolmm (NBT) stock solutton (stable at 4°C for at least 1 yr) dls- solve 0 5 g of NBT m 10 mL of 70% DMF m water

5 Bromochloromdolyl phosphate (BCIP) stock solution (stable at 4°C for at least

1 yr) dissolve 0.5 g of BCIP (dlsodmm salt) m 10 mL of DMF

6 Alkaline phosphatase buffer (stable at 4°C for at least 1 yr) 100 ti NaCl, 5 mM MgCl,, 100 mM Trrs-HCl (pH 9.5)

7 Enzyme substrate solutron add 330 PL of NBT stock solution to 50 mL of alka- lme phosphatase buffer Mix well and add 165 uL of BCIP stock solutton Use wtthm 1 h

8 Stop solution 20 mM EDTA m PBS (phosphate buffered salme)

2.72 Defection of Ligand Binding (Chemiluminescence Method)

England) 1: 10 m T-TBS and add 10% (w/v) sucrose

which IS directed agamst the primary ltgand 1 10,000 m blockmg buffer (see Note 3)

4 Detection reagent Just before developmg prepare the detectton reagent (BM Chemtlummescence Western Blottmg Reagents, Boehrmger Mannhelm, Mannhelm, Germany, or ECL Western Blottmg Detection Reagents, Amersham Buchler, Braunschwetg, Germany) according to the given protocols

5 Standard X-ray film and ftlmcassette, developer, water bath, and fixing solution

2.13 Regeneration of Cellulose-Bound Pepfide Libraries

1 Water

2 DMF

3 Regeneration buffer A drssolve urea (480 5 g) and sodturn dodecyl sulfate (10 0 g) m water (800 mL) Make up to 1 L with water, then add 1 mL of 2-mercaptoethanol

4 Regeneration buffer B: mix water (400 mL) with ethanol (500 mL) and add ace- tic acid (100 mL)

3 Methods

In Subheadings 3.1,3.6 we describe the manual synthesis of a combinato-

about the automated synthesis of peptide libraries using the Spot synthesizer Auto-Spot Robot APS 222 of Abimed GmbH In Subheadings 3.8 and 3.9 the synthesis of a mutational analysis and a peptide scanning library is described The screening methods of these different peptide libraries are explained in Subheadings 3.10,3.12 Subheading 3.13 describes the regen- eration of cellulose-bound peptide libraries

Cellulose Membrane Supports 37 3.1 Modification of the Cellulose

1 Mark 400 points as a 20 x 20 matrix on a square piece of cellulose paper (about

20 x 20 cm) usmg a penctl (graphite IS stable against the pepttde synthesis chemr- cals) For regular spotting, be sure that the points are visible from the other stde

of the paper

2 Incubate the dry cellulose paper with 12 mL of activated p-alanme solution for

3 h m a sealed metal or glass vessel without shaking Avoid air bubbles This IS done to mtroduce suitable anchor functions for the subsequent pepttde synthesis The activated p-alanine forms an ester bond with the hydroxyl groups of the cellulose

3 Wash the membrane three times with 50 mL of DMF for about 3 mm each A

should be carried out as described here

4 Cleave the Fmoc protecting groups by treatment of the paper with 50 mL of pip- ertdine solutton for 20 mm

5 Wash the membrane five times with DMF

6 Wash the membrane twice with 50 mL of methanol for 3 mm All subsequent methanol washing procedures should be carried out as decrtbed here

7 Dry the membrane

3.2 Definition of the Spots

1 For the second couplmg step, spot 1 uL of Fmoc-P-alanme-OPfp solutron to the predefmed positrons on the cellulose membrane Use the nonmarked stde of the paper Spot at the transparent pencil points A multtstep pipet is recom- mended Define an addmonal spot somewhere on the edge of the membrane This spot will be cut out m order to define the functtonalrty of the spots (see

Subheading 3.3.) After 15 mm reaction time repeat the spotting once to assure a complete coupling (15 mm reaction time)

2 Position the membrane carefully face-down m 20 mL of acetanhydrlde solution

A Avoid shaking and an- bubbles After 2 mm, incubate the membrane face up wrth 50 mL of acetanhydride solution B for 30 mm with shaking This 1s done to acetylate the amino functtons of the membrane that did not react with the second p-alanme Thus, defined sites for the following synthesis of the peptlde library can be achieved

3 Repeat steps 3-6 of Subheading 3.1

4 Stain the membrane with 50 mL of BPB solution A The BPB solutton should remam yellow and the spots should become blue, owmg to the basic character of the ammo groups of the coupled second p-alanme Treat the membrane untd an equal blue staining of the spots IS reached

5 Wash the membrane wrth methanol for 3 mm

6 Dry the membrane At this point the synthesis can be interrupted The membrane should be stored u-r a sealed plastic bag at -20°C until the next working day

32 Kramer and Schneider-Mergener 3.3 Functionality Determination

The randomized positions of a combinatorial peptide hbrary XXB,B,XX are introduced by double coupling an equimolar activated amino acid mixture

at 1.5 Eq in proportion to the amino functions (2) This is done to overcome the strong bras of the standard coupling conditions that reflect the different cou- pling rates of each amino acid Therefore, the number of free amino functions per spot has to be determined This does not need to be done for the synthesis

of peptide scanning libraries and mutational analyses, since these libraries do not contain randomized positions

I Cut out the addrtronal spot (see Subheading 3.1., step 1) with a hole puncher

2 Stam the spot with 1 mL of BPB-solutron B m a 1 5-mL tube untrl saturation

3 Wash three times with about 1 mL of methanol each

4 Dry the spots

5 Destam the spot completely with 1 mL of prperrdme solutron

6 Determine the extinction photometrrcally at a wavelength of 605 nm

7 Calculate the amount of ammo functrons on the spot using an extmctron coeffr- cient of &6a5 = 95,000 L mol-’ cm-‘, Presume that each ammo function bmds one bromophenol blue molecule Usually the membrane carrres about 60 nmol ammo functrons per 0 25 cm’-spot (1 pL creates a soaked area of about 0.25 cm*)

3.4 Coupling of the Amino Acids

The synthesis of cellulose bound peptides IS done as a cyclic process of double coupling the amino acids, washing with DMF, cleaving the Fmoc pro- tecting groups, washing with DMF and methanol, staining, washing with methanol, and drying After that, the synthesis of the next position of the peptrdes can be carried out Work according the synthesis diagram in Fig 1

Peptides are always synthesized from the C-terminal to the N-terminal posi-

tion, i.e., for the lrbrary XXB,B,XX, the synthesis steps are: X + X + Bz +

B, +X+X

1 For coupling randomrzed posrtrons X prpet I pL of the X-mixture onto each spot twrce (2 x 15 min reactron time) The couplmg reactron can be followed by a color change from blue to blue-green

2 Repeat steps 3-5 of Subheading 3.2 After that, the membrane IS ready for cou- pling the next position of the peptrde library (see Note 2)

3 For coupling the B-posrtrons, spot one of the 20 ammo acid solutions twice onto each column (for B2) or each row (for B,) of the 20 x 20 matrrx twice Use 1 VL per spot (2 x 15 mm reactron time) For a clear bmdmg analysis, an arrangement

in alphabetical order IS recommended, 1 e., spot A onto the first column/row and

Y onto the 20th column/row Coupling reactions can be followed by color change from blue to blue-green for Pfp esters and yellow for Dhbt esters

Cellulose Membrane Supports 33

4 After coupling the last position of the peptldes continue with steps 3-5 in Subheading 3.1 Do not stam the membrane

3.5 Acetylation of the N-Terminus

1 Incubate the membrane with 50 mL of acetanhydride B solution for 30 mm

2 Wash the membrane five times with 50 mL of DMF for 3 min

3 Wash the membrane twice with 50 mL of methanol

4 Dry the membrane

3.6 Cleavage of the Side Chain-Protecting Groups

1 Incubate the membrane with 100 mL of deprotecting solution m a tightly closed box for 2 5 h After a few seconds the cleaved “tntyl groups” of the cysteme row and column can be observed as yellow spots After treatment with the deprotection solution the membrane is much less mechamcally stable Therefore, handle the membrane extremely carefully from this point on

2 Wash 4 times with 50 mL of DCM for 3 mm

3 Wash three times with 50 mL of DMF for 3 mm

4 Wash twice with 50 mL of methanol for 3 mm

5 Dry the membrane

6 Store the membrane at -20°C

3.7 Automated Spot Synthesis

The disadvantages of manual synthesis are clear: the high number of synthe- sis steps are time consuming, the required precision during the pipeting pro- cess is not always manually managable Furthermore the error quote increases with the amount of peptides on the membrane The laborious pipeting proce- dure of the peptide library synthesis process led to the development of an auto- mated spot-synthesizer, which allows a maximum amount of precision and reliability for the synthesis The system Auto-Spot Robot APS 222 was devel- oped by Abimed GmbH especially for the synthesis of peptides on continuous membrane supports and automates the distribution of the amino acid deriva- tives The membrane can be divided into several trays and up to 8000 spots can

be synthesized simultaneously This allows the synthesis of a combinatorial peptide library with three defined positions The system only automates the pipeting procedures, all other steps, such as membrane modification, acetyla- tion, washing, Fmoc-deprotection, and side-chain deprotectlon have to be done manually

The software DIGEN, which can be purchased from Jermi BioTools GmbH, 1s recommended for the generation of sequence files of several types of peptide libraries, such as combinatorial libraries, peptide scanning libraries, mutational analyses of peptides, loop libraries, and so forth These files can be easily loaded into the Spot-synthesis software of the Auto-Spot Robot APS 222

34 Kramer and Schneider-Mergener

3.8 Synthesis of a Mutational Analysis

For the manual synthesis of a mutational analysis of a linear peptide use a matrix of 21 x peptide length (Fig 1) On the first column synthesize the ongi- nal peptide, on the other columns substitute each position of the peptide by all

20 amino acids m alphabetical order Carry out the peptide synthesis steps as described in the previous sections Here are some recommendations for the pipeting order

1 In the first coupling step, ptpet the C-terminal ammo acid of the original peptide

onto all rows except the last one

2 Ptpet the same ammo acid onto the first spot of the last column

3 Spot all 20 ammo acids successtvely onto the remaining 20 spots m alphabettcal order

4 Use an analog procedure for the remaining synthesis steps, 1.e , ptpet the subsmutton ammo acids m the second round onto the second row from below, and so on

3.9 Synthesis of a Pepticfe Scanning Library

The first step is the derivation of the peptrde sequences from the protein sequence A peptide length of 12 amino acids and an overlap of 9 amino acids

is recommended In Fig 2 the synthesized peptide sequences derived from the transforming growth factor-a sequence are given If you choose a higher over- lap, the spot number increases, but the information about the minimal hgand binding protein regions is more subtly differentiated The software DIGEN

generates sequence lists of peptide scanning libraries

The manual synthesis of a peptide scanning library requires a much higher degree of concentration than the synthesis of a combmatorial peptide library or

a mutational analysis owing to the irregular order of pipeting steps in each synthesis cycle To avoid mistakes, the use of a punch card, which covers the membrane during the pipeting, is recommended Every hole has to be labeled with the amino acid for the respective spot This has to be done for each syn-

thesis cycle

3.70 Screening of the Peptide Libraries

The screening strategy of the different peptrde libraries depends on various preconditions Is the hgand of interest available m an enzyme-labeled or radio- actively labeled form? Is an anti-ligand antibody available in an enzyme- labeled or radioactively labeled form? Is the ligand-peptide mteraction expected to have a high or low affinity? In this section, we describe the screen- ing of peptide libraries with a nonlabeled ligand If you have a labeled ligand available use the Incubating conditions of this section and the developing con- ditions of one of the next sections depending on your label If you do not

Cellulose Membrane Supports 35

have a labeled antiligand antibody available, you have to use a third labeled antibody In this case use the same incubating and washing conditions for the second and third antibody

Rinse the membrane with a small volume of methanol for 1 mm This IS done to avoid the precipitation of hydrophobic peptides during the following TBS wash- ing procedure

Wash the membrane three times with an appropriate volume of TBS for 10 mm The volume depends on the library and the vessel size The membrane should be sufficiently covered with the solutron

Block the membrane with the same volume of blocking buffer for about 14 h at

room temperature with shaking

Wash the membrane once with the same volume of T-TBS for 10 mm

Incubate the membrane with the same volume of hgand solution for 3 h (see

Note 3)

Wash the membrane three times with the same volume of T-TBS for 10 mm

3.11 Detection of Ligand Binding (Alkaline Phosphatase Method)

For the detection of high affinity bmdmg peptrdes, such as linear epitopes identified via peptide scanning libraries or mutational analyses of linear anti- body epitopes, the alkaline phosphatase method is recommended The BCIP/ NBT substrate generates an intense black-purple precipitate at the site of enzyme binding The reaction proceeds at a steady state rate, thus allowing control of the development of the reaction

1 Incubate the membrane with an appropriate volume of alkaline phosphatase- labeled antibody solutton for 2 h with gentle agitation The volume depends on the library and the vessel size The membrane should be suffictently covered

with the solution (see Note 3)

2 Wash three times with the same volume of T-TBS for 10 mm each

3 Incubate the membrane with the same volume of enzyme substrate solution Develop the membrane with agitation until the spots are suitably dark (l-30 mm)

4 Stop the reaction by rinsing the membrane with the same volume of stop solution three times for 3 min each The EDTA in the stop solution chelates the Mg2+ ions,

which are essential for alkalme phosphatase activity (see Note 4)

3.12 Detection of Ligand Binding (Chemiluminescence Method)

For the detection of low affinity binding peptides, such as peptide mixtures

in a library XXB ,B,XX, the chemiluminescence method is recommended The chemiluminescence method is a highly sensitive detection method, has short exposure times ranging from a few seconds to 1 h, and avoids the use of radio- activity The lummiscence reaction reaches its maximum after l-2 min and is relatively constant for 20-30 min After 1 h the signal intensity decreases to about 60-70% of maximum

36 Kramer and Schneider-Mergener

1, Incubate the membrane with an appropriate volume of peroxldase-labeled antl- body solution for 2 h with agitation The volume depends on the library and the vessel size The membrane should be sufflclently covered with the solution (see Note 3)

2 Wash three times with the same vol of T-TBS for 10 mm each

3 The following steps should be carried out m a dark room* incubate the membrane with 75 pL/cm2 of detection reagent for about 1 min Make sure that each part of the membrane 1s covered with detection solution

4 Insert the membrane protein side up mto a film cassette that 1s lined with a plastic film Cover the blot with a transparent plastic film

5 Switch off the light, place a sheet of film onto the membrane, and close the cas- sette Expose for 60 s

6 Immediately replace the exposed film with a new one, reclose the cassette, and develop the exposed film at once The developing process can be followed by using red safelights

7 Expose the second film for a suitable time (up to 1 h) estimated from the signal intensity on the first film (see Note 5)

3.13 Regeneration of Cellulose-Bound Pep tide Libraries

After probing, the cellulose peptide libraries can be regenerated using the following procedure The membrane can then be reprobed with the same or with a different ligand of interest With care the membrane may be regenerated and reprobed several times The success of the regeneration should be checked (see Note 6) Do not allow the membrane to dry out before regeneration

1 Wash the membrane three times with water for 10 mm The volume used for this and the followmg steps should be the same as m Subheading 3.10., step 1 or Subheading 3.11., step 1

2 Wash the membrane three times with DMF for 10 mm Prolong the DMF wash-

mg time, if the dye cannot be removed during the given time

3 Wash the membrane twice with water for 10 mm

4 Incubate the membrane with regeneration buffer A for 10 mm Repeat this step twice

5 Incubate the membrane with regeneration buffer B for 10 mm Repeat this step twice

6 Wash the membrane twice with methanol for 10 min

7 If the membrane 1s not immediately reprobed, then dry the membrane and store it

at -20°C

4 Notes

1 The purity of DMF and NMP is a critical point m peptide synthesis, for both DMF and NMP can degrade to give free ammes These ammes lead to premature Fmoc-deprotection or decomposltlon of ammo acid active esters This ~111 reduce

Cellulose Membrane Supports

the yield of full length peptide and cause byproduct formation, thus, only DMF and NMP free of amines should be used To test the purity add 10 ltL of 1% bromophenol blue solution m DMF to 1 mL of NMP or DMF in an 1 S-pL tube and mix thoroughly Allow to stand for 5 mm and then observe the color yellow indicates satisfactory to use, yellow/green or green/blue means do not use, pur- chase a better one

2 Bromophenol blue staining (Subheading 3.2., step 4) the Intensity of stainmg varies depending on the last spotted ammo acid Some ammo acids, such as cys- teine, aspartrc acid, glutamic actd, and asparagine stain only weakly Alanme, glycine, and prolme stam more strongly than others These differences may serve

as an internal control for correct pipeting

3 The incubation and washing condittons for the ligand of interest (Subheading 3.10.) and the second antibodies (Subheadings 3.11 and 3.12.) have to be opti- mized for each system In general, if you do not get any srgnal or your signals are too weak there are several posstbilities to optimize:

a Increase the ligand and/or antibody concentrations

b Prolong the incubation time with ligand to overnight at 4°C

c Prolong the mcubation time with secondary antibody to 3 h

d Shorten the washing times, use washing buffer without Tween-20

e Shorten the blockmg time to 3 h

If the background IS too high, try the followmg changes:

a Increase the detergent concentration m washing buffer

b Increase the washmg times and/or the washing volumes

4 Cysteme-containing peptides sometimes cause a signal m the alkaline phos- phatase detection system (Subheading 3.11.), which IS not a result of bmdmg to the conjugated antibody, but probably is caused by a catalytic reaction of the thiol group of the cysteme side chain with the bromochloroindolyl phosphate substrate, which forms the product If you get those spots colored, remember this reaction and synthesize on your next membrane control peptides, m which cys- teine is substituted by the physrcochemically related ammo acid serme Another possibility to circumvent this problem is to use the chemiluminecence detection method

5 In the chemiluminescence detection method (Subheading 3.12.) somettmes clear spots on a black background appear In this case the ligand and/or secondary antibody concentrations are much too high On the spots, which carry a high amount of antibody conjugate, all the substrate is used up before the X-ray film can be placed on the membrane, resulting in clear spots Wash extensively with

T-TBS and try to redetect, or regenerate the membrane and incubate with lower concentrations of proteins

6 The success of the regeneration should be checked after each regeneration by incubating with only the secondary antibody and subsequent detection If you have used a directly labeled ligand, repeat the detection procedure after regenera- tron In some cases very strongly bmdmg proteins cannot be removed from the membrane

1 Frank, R (1992) Spot synthesis an easy technique for the positionally address-

Kramer, A , Schuster, A , Remeke, U., Maim, R , Volkmer-Engert, R., Landgraf,

C , and Schnetder-Mergener, J (1994) Combmatorial cellulose-bound peptide libraries: screening tool for the identification of peptides that bmd hgands with predefmed specificity Methods f&388-395

Remeke, U., Sabat, R , Kramer, A., Stigler, R -D., Seifert, M , Michel, T., Volk, D., and Schneider-Mergener, J (1995) Mapping protem-protein mteractions using hybritope and peptide scanning libraries Mol Dzverszty 1,141-148

Schneider-Mergener, J., Kramer, A , and Remeke, U (1996) Peptide libraries bound to contmuous cellulose membranes tools to study molecular recognition,

m Combinatorial Llbrarles, (Cortese, R , ed.), W de Gruyter, Berlin, pp 53-68 Kramer, A , Vakalopoulou, E., Schleunmg, W.-D , and Schneider-Mergener, J (1995) A general route to fingerprint analyses of peptide-antibody mteractions using a clustered ammo acids peptide library* comparison with a phage display library Mol Immunol 32,459-465

cellulose-bound peptide libraries for the detection of antibody epitopes, m Pep- tides 1994 (Mala, H L S., ed ), ESCOM, Leiden, pp 475,476

Stigler, R., Rdker, F , Katmger, D., Elliot, G., Hohne, W , Henklem, P , Ho, J X , Kramer, A , Nugel, E , Porstmann, T , and Schneider-Mergener, J (1995) Charac- terization of the Interaction between a Fab fragment against gp4 1 of HIV- 1 and Its peptide epitope using a peptide epitope library and molecular modellmg Protezn Eng $471479

Volkmer-Engert, R., Ehrhard, B , Hohne, W , Kramer, A., Hellwig, J., and Schneider-Mergener, J (1994) Preparation, analysis and antibody bmdmg studies

of simultaneously synthesized soluble and cellulose-bound HIV- 1 p24 peptide epttope libraries Lett Pept Sci 1,243-253

Weiergraber, 0 , Schneider-Mergener, J , Grdtzmger, J , Wollmer A., Kuster, A., Exner, M., and Hemrich, P C (1996) Use of immobilized synthetic peptides for the identification of contact sites between human mterleukm-6 and its receptor FEBS Lett 379,122-126

38

Acknowledgments

Kramer and Schneider-Mergener

This work was supported by the BMBF, DFG, and Fonds der Chemlschen Industrie We thank Beret Hoffmann and Christlane Landgraf for excellent tech- nical assistance

References

Cellulose Membrane Supports 39

11 Riidtger, S , Germeroth, L., Schneider-Mergener, J., and Buckau, B (1997) Sub- strate spectfrcny of the DnaK chaperone determined by screening cellulose-bound pepttde libraries EMBO J 16,1501-1507

12 Malm, R , Steinbrecher, A , Semmler, W., Noll, B , Johannsen, B., Frommel, C , Hohne, W., and Schneider-Mergener, J (1995) Identification of technetium-99m binding peptides usmg cellulose-bound combmatortal peptide libraries J Am Chem Sue 117,11,821-l 1,822

13 Wmkler, D., Schuster, A , Hoffmann, B., and Schneider-Mergener, J (1995) Synthesis of cyclic pepttde libraries bound to contmuous cellulose membrane supports, m Peptzdes 1994 (Mata, H L S., ed ), ESCOM, Letden, pp 485,486

14 Wmkler, D , Stigler, R , Landgraf, C., Hellwig, J , and Schneider-Mergener, J (1995) Determination of the bmdmg conformations of peptide epttopes usmg cyclic peptide libraries m, Peptzdes 1995 (Kaumaya, P T P and Hodges, R S., eds.), ESCOM, Leaden, pp 315,316

15 Geysen, H M , Rodda, S J , and Mason, T J (1986) A prlorl delmeation of a peptide which mimics a discontmuous anttgemc determinant Mol Immunol 23, 709-715

16 Houghten, R A , Pmilla, C , Blondelle, S E , Appel, J R , Dooley, C T , and Cuervo, J H (1991) Generation and use of synthettc peptide combinatorial librar- ies for baste research and drug discovery Nature 354,84-86