Non-proline Cis Peptide Bonds in Proteins

In peptides and proteins amino acids are linked together via their carboxylate carbon and their amine nitrogen atoms. Due to a partial double bond character of this peptide bond between C and N, the rotation around this bond is restricted and only two conformations are energetically preferred . The conformation, in which the dihedral angle omega is 180 deg. is called the trans conformation, and the one, in which omega is 0 deg. is called the cis conformation [IUPAC, J.Mol.Biol. 1970, 52, 1-17].

In protein structures, the peptide bond conformation is found to be trans in the absolute majority of the cases (Ramachandran & Sasisekharan, 1968). An exception to this constitute the peptide bonds between any amino acid and Pro (Xaa-Pro), for which an appreciable fraction occurs in the cis conformation. A survey by Stewart et al. (1990) found only 0.05% of all Xaa-nonPro, but 6.5% of all Xaa-Pro peptide bonds to occur in the cis conformation, MacArthur and Thornton (1991) found 5.5% cis for Xaa-Pro and recently we (Weiss et al., 1998a) found in a much larger non-redundant set of 571 proteins 5.2% cis for Xaa-Pro and 0.03% cis for Xaa-nonPro.

The reason for this preponderance of the trans peptide bond is believed to lie in the energy difference between the two different forms. It is clear that the cis form is energetically less stable due to a steric repulsion of the two neighbouring Calpha atoms, but the absolute numbers have been the basis of much debate over the years. Experimentally, Radzicka et al. (1988) found that the model compound N-methylacetamide occurs at about 1.5% in the cis form, regardless of the solvent. From this, a difference in free enthalpy of 2.4 kcal/mol can be computed. Almost the same value was reported by Drakenberg and Forsén [J.Cem.Soc.,Chem.Comm. 1971, 1404-1405]. They found deltaG = 2.5 (+/- 0.4 kcal/mol in water. For Pro-containing peptides, Grathwohl and Wüthrich (1976) reported a cis content of about 10 - 15% when the charge at the C-terminus is removed either by protonation or by a protecting group. This corresponds to a free enthalpy difference between cis and trans of about 1.0 - 1.3 kcal/mol at 293 K.

Theoretical calculations are pretty much in accord with these experimental values. Maigret et al. (1970) reported a 0.5 kcal/mol difference between cis and trans for Acetyl-Pro-N(CH3) based on quantum mechanical calculations, and Jorgensen and Gao [J.Am.Chem.Soc. 1988, 110, 4212-4216] reported 2.5 kcal/mol difference for the model compound N-methylacetamide in the gas phase and 2.6 kcal/mol in water. Since these numbers depend critically on the geometrical and force field parameters used, they have to be taken with care.

In proteins, the energetic situtation is more difficult to describe and therefore less clear. Ramachandran and Mitra (1976) used conformational energy calculations of tripeptide units to derive expected frequencies of 0.1% and 30% cis for an Ala-Ala and an Ala-Pro peptide bond respectively. These numbers correspond to respective enthalpy differences of 4.0 kcal/mol and 0.5 kcal/mol.

Experimental data for proteins stem mainly from cis-Pro point mutations. Tweedy et al. (1993) found that the Pro202->Ala mutant of carbonic anhydrase is by about 5 kcal/mol less stable than the wild type enzyme. In both the wild type and the mutant enzyme the peptide bond between residues 201 and 202 oocurs in the cis conformation; therefore the decrease in stability of the protein upon mutating Pro202 to Ala is thought to be mainly due to the less favorable cis/trans equilibrium, although the authors say that other factors may also contribute. Schultz and Baldwin (1992) reported a destabilization of 2.7 kcal/mol for the Pro93->Ala mutant of ribonuclease A with respect to the wild type protein. In the three-dimensional structure of this mutant (Pearson et al., 1998) the loop containing the Tyr92-Ala93 cis peptide bond becomes more mobile, but it seems as if the cis conformation is retained. Mayr et al. (1993) also found a very strong destabilization of about 5 kcal/mol for the Pro39->Ala mutant of ribonuclease T1, but in this case it's not clear whether the mutated protein still contains the peptide bond in the cis conformation.

Due to the double bond character of the amide bond, an appreciable barrier exists for the rotation around the C-N bond. Experimental results for the activation enthalpy for model compounds (Drakenberg & Forsén, 1971) and theoretical calculations ([Christensen et al., J.Chem.Phys. 1970, 53, 3912-3922]; [Perricaudet & Pullman, 1973]) agree on values of about 20 kcal/mol for the activation enthalpy. Such a large barrier makes the interconversion between the two conformations a rather slow process at room temperature. If one assumes for simple stereochemical reasons that all peptide bonds are synthesized in the same conformation, which must be obviously the trans conformation, then it becomes clear that isomerization must occur at some stage of the folding process of the protein, if the correctly folded protein contains at least one peptide bond in cis conformation. And indeed it has been demonstrated (Brandts et al., 1975) that isomerization of the peptide bonds before proline residues plays a decisive role in protein folding. The discovery of prolyl-cis/trans-isomerases (Fischer et al., 1984) supports this notion. It has been shown that these enzymes catalyse the cis/trans isomerization of Xaa-Pro bonds, but not of Xaa-nonPro bonds (Scholz et al., 1998a). Recently, a ribosome-associated prolyl isomerase named trigger factor has been described (Stoller et al., 1995) which binds unfolded proteins that do not contain proline residues (Scholz et al., 1998b). However, isomerization does not take place (Scholz et al., 1998a) and it remains unclear, what happens to the non-Pro cis peptide bonds during the course of folding.

The occurrence of non-Pro cis peptide bonds has been associated with steric strain in proteins (Herzberg and Moult, 1991) similar to the occurrence of residues with unfavorable phi/psi-angles and it has been speculated that the location of these cis peptide bonds is often a peculiar one with respect to the function of the molecule (Stoddard et al., 1998; Weiss et al., 1998). It has been discussed that these sites of strain are some kind of energy reservoir for the protein. In the course of a chemical reaction or a conformational change the energy that could be liberated by a conversion of a cis peptide bond to the trans conformation could help drive the reaction towards the product. This notion, however, is speculative at the time and has to await further experimental confirmation.

The Database

Amino Acid Distribution in Proteins

Cis Peptide Bonds in Proteins - a General Statistic

	All	<2.0 Å	2.0 Å - 2.5 Å	2.5 - 3.5 Å
Number of Proteins	571	291	184	96
Number of peptide bonds	153209	72567	52194	28448
Xaa-Pro	7413	3407	2566	1440
Xaa-non Pro	145796	69160	49628	27008
cis peptide bonds	429	232	142	55
	(0.28%)	(0.32%)	(0.27%)	(0.19%)
Xaa-Pro	386	205	129	52
	(5.21%)	(6.02%)	(5.03%)	(3.61%)
Xaa-non Pro	43	27	13	3
	(0.029%)	(0.039%)	(0.026%)	(0.011%)

4.8% Xaa-Pro (4.7% Stewart et al. J.Mol.Biol. 214(1990)253)
95.2% Xaa-non Pro

0.3% of all peptide bonds found to be in cis

90% Xaa-Pro (about 5% of the total)
10% Xaa-non Pro (about 0.03 % of the total)

At high resolution, the number of Xaa-Pro cis peptide bonds is about twice as high than at medium
and low resolution and the number of Xaa-non Pro bonds in cis conformation is about four times
as high.

Non-proline Cis Peptide Bonds - Distribution in the Database

Geometry Geometry of Cis Peptide Bonds in Proteins

Bond	Value [Å]¹	$\sigma$	Value [Å]²	$\sigma$	Value [Å]³	$\sigma$ ⁴	Value [Å]⁵	$\sigma$
N-C $_{\alpha}$	1.458	0.021	1.458	0.019	1.488	0.008	1.460⁶	0.019
C $_{\alpha}$ -C	1.527	0.017	1.525	0.021	1.508	0.009	1.527	0.025
C-O	1.236	0.016	1.231	0.020	1.244	0.008	1.238	0.013
C-N⁺	1.329	0.016	1.329	0.014	1.376	0.007	1.336	0.010
N⁺-C $_{\alpha}^{+}$	1.456	0.013	1.458	0.019	1.457	0.009	1.459	0.007
Angle	Value $[^{\circ}]$ ¹	$\sigma$	Value $[^{\circ}]$ ²	$\sigma$	Value $[^{\circ}]$ ³	$\sigma$ ⁴	Value $[^{\circ}]$ ⁵	$\sigma$
N-C $_{\alpha}$ -C	109.2	4.0	111.2	2.8	106.7	0.6	108.6⁶	3.0
C $_{\alpha}$ -C-O	119.1	2.9	120.8	1.7	121.1	0.5	119.7	1.3
C $_{\alpha}$ -C-N⁺	120.3	5.5	116.2	2.0	118.3	0.6	119.7	1.2
C-N⁺-C $_{\alpha}^{+}$	126.8	4.9	121.7	1.8	127.5	0.6	127.8	1.4
O-C-N⁺	120.3	4.3	123.0	1.6	120.3	0.6	120.6	1.1
N⁺-C $_{\alpha}^{+}$ -C⁺	109.9	4.3	111.2	2.8	114.5	0.6	112.9	2.6

¹ from 27 non-Pro cis peptide bonds in proteins determined at a resolution of $\leq$ 2.0 Å
² Engh R.A & Huber, R. Acta Cryst. 1991, A47, 392-400
³ from Ala-Asp cis peptide bond in 0.94 Å ConA structure [Deacon et al., J. Chem. Soc.
Faraday Trans. 1997, 93, 4305-4312]
⁴ Experimental standard deviations from restrained refinement
(Dr. A. Deacon and Prof. J. Helliwell, personal communication.)
⁵ from 16 small molecule entries retrieved from the Cambridge Structural Database.
⁶ based on 7 data points only, since not all the amide bonds from the CSD are in peptides.

Proteins Containing Non-proline Cis Peptide Bonds

PDB Code	Protein	Resolution¹	Non-Proline Cis Peptide Bonds²	Additional Cis Peptidyl-Prolyl Bonds	Author(s)¹
1AMP	aminopeptidase	1.8	D117-D118		Chevrier et al.
1BMF	mitochondrial F1-ATPase	2.85	D269-D270 (A)	+	Abrahams et al.
			D256-N257 (D)
1CEC	endoglucanase CelC	2.15	W313-N314	+	Alzari & Dominguez
1CLC	endoglucanase CelD	1.9	D177-A178	+	Alzari & Lascombe
1CTN	chitinase A	2.3	G190-F191		Perrakis et al.
			E315-F316
			W539-E540
1DYR	dihydrofolate reductase ( P. carinii)	1.86	G124-G125	+	Champness et al.
1ECE	endocellulase E1	2.4	W319-S320	+	Sakon et al.
1F13	coagulation factor XIII	2.1	R310-Y311	+	Weiss & Hilgenfeld
			Q425-F426
1GAI	glucoamylase-II	1.7	G23-A24	+	Aleshin et al.
1GHR	1,3-1,4- $\beta$ -glucanase	2.2	F275-A276	+	Varghese & Garrett
1GSA	glutathione synthetase	2.0	V113-N114	+	Hara et al.
1HGX	H-G-X phosphoribosyltransferase	1.9	L46-T47		Somoza et al.
1JAP	matrix-metalloproteinase-8	1.82	N188-Y189		Bode et al.
1JPC	snowdrop lectin (mannose-specific)	2.0	G98-T99		Wright & Hester
1LEN	lentil lectin	1.8	A80-D81		Van Overberge et al.
1LUC	bacterial luciferase	1.50	A74-A75	+	Fisher & Rayment
1MHL	myeloperoxidase	2.25	N549-N550 (C)	+	Fenna et al.
1MKA	$\beta$ -hydroxydecanoyl ACP dehydrase	2.0	P31-N32		Leesong
			H70-F71
1NAR	narbonin	1.8	G38-F39	+	Hennig et al.
			W261-N262
1NBA	N-carbamoylsarcosine amidohydrolase	2.0	A172-T173		Romao et al.
1ORO	orotate phosphoribosyltransferase	2.4	A71-Y72		Henriksen et al.
1PBG	6-phospho- $\beta$ -D-galactosidase	2.3	W421-S422	+	Wiesmann & Schulz
1PGS	N-glycosidase F	1.8	C204-A205	+	Norris et al.
1TPL	tyrosine phenol-lyase	2.3	V182-T183	+	Antson et al.
1XYZ	endo-1,4- $\beta$ -xylanase Z	1.4	H596-T597		Alzari et al.
1ZQA	DNA polymerase $\beta$	2.7	G274-S275		Pelletier & Sawaya
2CTC	carboxypeptidase A	1.4	S197-Y198		Teplyakov et al.
			P205-Y206
			R272-D273
2EBN	endoglycosidase F1	2.0	F45-S46	+	Van Roey
2HVM	hevamine	1.80	A31-F32	+	Van Scheltinga et al.
			W255-S256
2MAD	methylamine dehydrogenase	2.25	K129-A130		Huizinga et al.
2REB	Rec A protein	2.3	D144-S145		Story & Steitz
2TMD	trimethylamine dehydrogenase	2.4	T70-H71		Mathews et al.
3DFR	dihydrofolate reductase ( L. casei)	1.7	G98-G99	+	Filman et al.
4AAH	methanol dehydrogenase	2.4	K269-W270	+	Mathews & Xia

¹ The resolution and the authors quoted are the ones that appear in the respective PDB entries.
² In case there are different polypeptide chains in the coordinate entry, the identifier of the
chain containing the cis peptide bond is given in parentheses. Cases with non-crystallographic
symmetry are only listed once and are not identified explicitly.

Non-proline Cis Peptide Bonds and Functional Implications

Side Chain Interaction in Non-proline Cis Peptide Bonds

The occurrence of aromatic residues involving a non-proline cis peptide bond is very high

aromatic residue N-terminal: 1CEC, 1CTN, 1ECE, 1GHR, 1NAR, 1PBG
2EBN, 2HVM

                                                              the interaction between the aromatic ring and
                                                              the side chain (especially the C(beta) atom)
                                                              is well defined (JPEG)

aromatic residue C-terminal 1CTN, 1JAP, 1NAR, 1ORO, 2CTC, 2HVM
4AAH

the strucural pattern is not well defined
(JPEG)

43 non-proline cis peptide bonds were found in the 25% database

10 non-proline cis peptide bonds with interaction between aromatic an aliphatic
side chains

6 non-proline cis peptide bonds with an N-terminal Trp residue

3 Trp-Ser cis peptide bonds in different proteins having different functions
with identical conformations (JPEG)

amino acid	25% database number %	from sequence ^a) %
Gly	12160 7.9	7.2
Ala	13120 8.6	8.3
Val	10507 6.9	6.6
Leu	13264 8.6	9.0
Ile	8724 5.7	5.2
Phe	6265 4.2	3.9
Tyr	5690 3.8	3.2
Trp	2313 1.5	1.3
Pro	7255 4.8	5.1
Cys	2080 1.4	1.7
Met	3267 2.2	2.4
Ser	9176 6.0	6.9
Thr	8940 5.9	5.8
Lys	8533 5.7	5.7
Arg	7310 4.8	5.7
His	3405 2.3	2.2
Asp	9114 5.9	5.3
Glu	9405 6.1	6.2
Asn	7028 4.7	4.4
Gln	5653 3.7	4.0

peptide bond	total number of occurences	number in cis conformation	frequency [%]
Xaa-aliphatic (except Pro) Xaa-polar Xaa-aromatic all	53629 74494 17673 145796	8 22 13 43	0.015 0.030 0.074 0.029
aliphatic-non-Pro polar-non-Pro aromatic-non-Pro all	58203 70906 16687 145796	17 16 10 43	0.029 0.023 0.060 0.029
aliphatic-Pro polar-Pro aromatic-Pro all	2860 3586 967 7413	152 157 77 386	5.31 4.38 7.96 5.21

metal binding
1AMP	Aminopeptidase	Asp-Asp	Zn⁺⁺

dimerization site
1BMF	F1 ATPASE	Asp-Asp
1HGX	Phosphoribosyltransferase	Leu-Thr
1MKA	Thiol ester hydrolase	His-Phe
1ORO	Phosphoribosyltransferase (Orotate)	Ala-Tyr
2TMD	Trimethylamine dehydrogenase	Thr-His

active site
1CTN	Chitinase A	Gly-Phe
		Glu-Phe
		Trp-Glu
1LEN	Lectin	Ala-Asp
1TPL	Tyrosine phenol-lyase	Val-Thr
2EBN	Endo-beta-N-acetyl-glucosamidase	Phe-Ser
3BLM	Beta-lactamase	Glu-Ile
5CPA	Carboxypeptidase	Ser-Tyr
		Arg-Asp

cofactor/substrate binding
1DYR	Dihydrofolatereductase	Gly-Gly	NADPH
1ECE	Endocellulase E1	Trp-Ser	Cellotetraose
1LLO	Hevamine	Ala-Phe	Allosamidin/
		Trp-Ser	Allosamizalon
1MKA	Thiolester hydrolase	His-Phe	3-Decynoyl-N-Acetyl-
			Cysteamine

General Aspects

The Database

Amino Acid Distribution in Proteins

Cis Peptide Bonds in Proteins - a General Statistic

Non-proline Cis Peptide Bonds - Distribution in the Database

Geometry Geometry of Cis Peptide Bonds in Proteins

Proteins Containing Non-proline Cis Peptide Bonds

Non-proline Cis Peptide Bonds and Functional Implications

Side Chain Interaction in Non-proline Cis Peptide Bonds

Conclusions

Further Reading