David Shaw King
HHMI/University of California at Berkeley
Frequently we are called upon to synthesize relatively large peptides, in the range of 50 to 100 residues, the size of an average protein domain. The majority of these peptides are used for structural analysisby electron paramagnetic resonance, NMR, and xray crystallographyor are immobilized as bait in affinity chromatography. The quantity required ranges from 10 to 100 mg, and the purity requirementsparticularly for xray crystallographyare high. Over the last several years we have synthesized several dozen large peptides, using Fmoc chemistry and ABI 431A synthesizers, and have employed several carboxyl-activation strategies. All syntheses generated the desired peptide, but the success of each synthesis has been a function of the amino acid sequence and of the synthesis strategy, which has improved every year. The purpose of this informal "how I make grandmother's ginger cookies" article is to illustrate and to pass along briefly various successful techniques of synthesis. The comparison of synthetic strategies is neither systematic nor exhaustive, and none of the techniques mentioned here is particularly sophisticated.
Fmoc chemistry has in our hands always produced peptides of higher quality and in greater yield than Boc chemistry. Pfp esters in the presence of DIEA or NMM yield large peptides of relatively poor quality (Pfp, pentafluorophenyl; DIEA, N,Ndiiso-propylethylamine; NMM, Nmethylmorpholine). We find that large peptides synthesized via HOBt esters give crude product of the best quality. HOBt esters formed by HBTU (in DMF, with DIEA or NMM) work well (see Figure 4) [HOBt, 1hydroxybenzotriazole; HBTU, (2(lHbenzo-triazol-1-1yl)1,1,3,3tetramethyluronium hexafluoro -phosphate]. Those formed by DCC (in NMPDCM), although a slower procedure, produce peptides that are noticeably cleaner (see Figure 5) (DCC, N,Ndicyclohexylcarbodiimide). The basic 0.1 mmol 431A HOBtDCC synthesis cycles devised by Ken Otteson at Perkin Elmer/Applied Biosystems Division are an excellent point of departure. Minor changes we have made for the synthesis of larger peptides are: (1) to increase the ratio of the acylating species over amine from 10:1 to 20:1 or even higher as the peptide is being synthesized, starting at cycle 4050, or whenever there is a stretch of vicinal bbranched amino acids or amino acids with bulky protecting groups close to the acarbon atoms; and (2) to acylate and deprotect for relatively long periods of time. This is important. We routinely deprotect the Fmoc group with 1M piperidine in NMP for 30 minutes, and acylate for 1 hour, increasing the acylation time to 2 or more hours as the peptide becomes longer. We have never found double coupling to be necessary.
NMP is the preferred solvent for acylation, deprotection, and rinses of polystyrenedivinylbenzene resins because the resins swell more in NMP, and hence enable better access of solvent and reagents to the growing peptide chain. Synthesis levels of 0.05 to 0.1 mmol yield quantities of final purified product, even with the most difficult syntheses, sufficient for NMR studies. We have not found resin substitution levels (range 0.15 to 0.95 mmol/g) to affect the quality of large peptides; what matters more is the quality of the resin.
Analysis of crude peptide by electrosprayionization mass spectrometry (ESIMS) in 0.1% TFA, 50% acetonitrile yields more information about the details of peptide synthesis and deprotection than any other single analytical method, and we use it routinely. Identification of deletions and incompletely deprotected species is straightforward, and the analysis is close to quantitative. Inspection of the ESI-MS spectra in Figure 1, showing a small peptide for clarity, reveals a phenomenon we see routinely: the efficiency of synthesis is always greater than the efficiency of deprotection. This is particularly true of peptides containing a long segment of residues with protecting groups close to the acarbon atoms; hindered tertbutyl esters and ethers and trityl groups (especially the tritylthioether on Cys) are the most tenacious protecting groups.
Figure 1: Crude 4l-residue peptide containing six Cys residues after three successive 6-hour cleavage/deprotections. The peptide sequence is GNCKC SGKPL TGYVD LGYCN EGWEK CASYY SPIAE CCRKK K. Left panels are electrospray mass spectra, right panels are reversedphase HPLC profiles. This illustrates a short peptide with tenacious protecting groups.
The cleavage/deprotection cocktail we have found to be the most effective for large peptides is reagent K, with a reaction time of 3 to 4 hours at room temperature. If deprotection is still incomplete, repeated deprotection of crude, dry peptide in fresh cleavage reagent for another 3 to 4 hours is necessary. Figure 1 shows one peptide after three consecutive deprotections. It is never a good idea to cleave a peptide in the same deprotecting mixture for a longer period of time, because the quality of peptide begins to deteriorate after 6 hours.
Figure 2: Crude 64-residue peptide after a single 4-hour cleavage /deprotection. The peptide sequence is LRLKI YKDTE GYYTI GIGHL LTKSP SLNAA KSELD KAIGR NTNGV ITKDE AEKLF NQDVD AAVR. Upper panel, reversedphase HPLC profile; lower panel, electrospray mass spectrum. This peptide presents no problems in synthesis or deprotection.
Figure 3: Electrospray mass spectra of a crude 54-residue peptide containing eight Cys residues after two successive 5 -hour cleavage/deprotections. The peptide sequence is EDKCK KVYEN YPVSK CQLAN QCNYD CKLDK HARSG ECFYD EKRNL QCICD YCEY. This illustrates a more difficult synthesis and deprotection.
Up to synthesis cycle 50 or 60 one rarely experiences any
synthetic difficulty. Figure 2 shows a crude, routine 64-residue
peptide after a single 3.5-hour deprotection, a relatively simple
synthesis. The purity of the resulting peptide was 85%, and the total
yield of isolated product at more than 97% purity was about 55%.
Figure 3 shows a considerably more difficult peptide, 54 residues
long with 8 Cys residues, deprotected twice.
Figure 4: Electrospray mass spectra of crude 69-residue (upper panel) and crude 85-residue (lower panel) peptides from the same synthesis, each after a single 4-hour cleavage/deprotection. The peptide sequence is [SPDFR IAFQE LLCLR R]/SSLA KAYGN GYSSN GNTGE QSGYH VEQEK ENKLL CEDLP GTEDF VGHQG TVPSD NIDSQ GRNCS TNDSLL. This illustrates how a short segment of hindered residues can lower synthesis quality. This peptide was synthesized by HBTU/HOBt activation; note the relatively poor quality of crude 85-residue peptide.
Figure 5: Electrospray mass spectra of a 78-residue peptide. The peptide sequence is QRLAP LSRKL FPVLY PQPSS NLERY LRSTF DEAEI KGDLL NQVLE KVNAA TETSQ KGSSA LKILT SLFNK CKKVN GGC. Upper panel, crude peptide after two 5-hour cleavage/deprotections; lower panel, product after a single reversed-phase HPLC purification. This illustrates a routine but relatively difficult synthesis.
Beyond cycle 50 or 60, deprotection and acylation cycles should be more conservative. The success of the synthesis depends on the sequence: some 85-mers can look better than some 55-mers. Figure 4 shows ESIMS of crude 69-mer and crude 85-mer from the same synthesis, each after a single 4 -hour deprotection. It is clear that the quality of the 69-mer is high, as the sequence up to this point is unhindered; between cycle 69 and 85, the sequence is highly hindered and the quality of the peptide suffers accordingly. The synthesis of a routine, relatively difficult 78-residue peptide is shown in Figure 5. The total synthesis took ten days. ESIMS analysis of the crude peptide after a single 4.5-hour deprotection shows a decidedly depressing looking product, but after two deprotections and a careful single step purification by reversed-phase HPLC, the correct peptide was obtained as an isolated product in 9% yield and with a purity of 96%, sufficiently high for analysis by NMR and xray crystallography.
We find that obtaining peptides of the 50- to 100- residue size range by synthesis is faster and cheaper than by cloning and overexpressing in E. coli, with the added bonus that the resulting peptide is unambiguously correct. With synthetic peptides, one does not have to deal with problems such as poor or no expression, cloning errors, cleavage of hemaglutinnin, Flag, or polyoma tags, metabolism of the His tag, aminoterminal processing, premature termination of translation, or mistranslation of nonpreferred codonsall problems attendant to the expression of eukaryotic genes in prokaryotic hosts.
The author may be contacted at the Howard Hughes Medical Institute/University of California at Berkeley, Department of Molecular and Cell Biology, Berkeley CA 94720, E-mail: dk@bosco2.berkeley.edu.
Return to the The ABRF Home Page