OUP user menu

A pepD-like peptidase from the ruminal bacterium, Prevotella albensis

Nicola D. Walker , Neil R. McEwan , R. John Wallace
DOI: http://dx.doi.org/10.1016/j.femsle.2004.12.032 399-404 First published online: 1 February 2005


Peptidases of Prevotella spp. play an important role in the breakdown of protein to ammonia in the rumen. This study describes a peptidase cloned from Prevotella albensis M384. DNA from P. albensis was used to complement a peptidase-deficient strain of Escherichia coli, CM107. A cloned fragment, Pep581, which enabled growth of E. coli CM107, contained an ORF of 1452 bp, encoding a 484 amino acid residue protein with a calculated molecular weight of 53.2 kDa and a theoretical pI of 4.90. Pep581 shared similar sequence identity of 47% with PepD from E. coli, and it was also a metallo-aminopeptidase. A putative catalytic metal binding region was identified in Pep581, similar to that found in the related PepT (a tripeptidase) and PepA (an oligopeptidase). Gel filtration indicated Pep581 was a dimer in its native state, similar to PepD of E. coli. PepD is a broad specificity dipeptidase that has been found in several prokaryotes. The enzyme expressed from Pep581 differed from PepD enzymes previously characterised in that it hydrolysed tri- and oligopeptides in addition to dipeptides, cleaving single amino acids from the N terminus.

  • Aminopeptidase
  • Prevotella albensis
  • Peptidase
  • Rumen

1 Introduction

Peptide breakdown in the rumen is part of a degradation process whereby dietary protein is broken down to ammonia, leading to the inefficient utilisation of dietary amino acids [1,2]. The genus Prevotella plays a significant role in this process [28]. The peptidases of Prevotella albensis M384 have been investigated biochemically [5,9,10], and one, a dipeptidyl peptidase IV (DPP-4), has been cloned and characterised [11]. Here, another peptidase, found by complementation of a peptidase-deficient strain of Escherichia coli, is described. Pep581 shares sequence similarity with other bacterial peptidase genes but has significant differences in substrate specificity.

2 Materials and methods

2.1 Cloning of peptidase from P. albensis

P. albensis M384 was isolated from the rumen of a sheep [12] and is maintained in the culture collection of the Rowett Research Institute. E. coli CM107 (Δleu-9lacZ521metthyA pepA11pepN102pepB1), a peptidase-deficient mutant [13], was a gift from Dr. C.G. Miller, University of Illinois, USA. Competent cells of E. coli were prepared using the method described by Hanahan et al. [14]. High-molecular-weight DNA was prepared from a 10-ml overnight culture of P. albensis grown on ruminal fluid-containing M2 [15]. Genomic DNA was isolated using Qiagen columns?, following the manufacturer's guidelines (Qiagen Ltd, Crawley, West Sussex, UK), and a plasmid genomic library was prepared in pBluescript. Competent cells of CM107 were transformed and recombinants were selected for their ability to grow in the presence of ampicillin (50 μg/ml). Clones were screened for peptidase activity against the synthetic substrate Ala2-β-naphthylamide (βNA) using an agarose overlay technique [13]. Untransformed CM107 was unable to hydrolyse this substrate. A resulting positive clone, Pep581, was characterised biochemically and genetically.

2.2 Measurement of peptidase activities

Clone Pep581 was grown overnight on LB + 50 μg/ml kanomycin medium at 37 °C with shaking. Cells were harvested by centrifugation (27,500g, 4 °C, 15 min), then washed in 50 mM potassium phosphate buffer, pH 7.0, and resuspended in 20 ml of the same buffer. The resuspended cells were sonicated at 30 μm in 30-s bursts on ice, with 30-s cooling intervals, for a total sonication time of 10 min using a Soniprep 150 (MSE Instruments, Crawley, Sussex, UK). Intact cells and cell debris were removed by centrifugation (27,500g, 4 °C, 15 min). The activity of the supernatant against a variety of different synthetic peptides was determined using chromogenic and fluorogenic substrates [10]. Reverse-phase HPLC was used to measure the breakdown of di-, tri and oligopeptides [9,16]. Inhibitors were investigated as described by Wang et al. [17]. Protein was determined by a modification of the Lowry method [18].

2.3 Gel filtration

The molecular mass of the native enzyme was estimated using gel filtration chromatography. Three milliliters of sonicated clone (0.25 mg protein/ml) was applied to a Sephacryl-300 column and eluted using 50 mM potassium phosphate buffer, pH 7.0, at 0.25 ml/min. Fractions were tested for their ability to hydrolyse GlyPro-4-methoxynaphthylamide (GlyPro-MNA), LysAla-MNA and GlyPro-p-nitroanilide (GlyPro-pNA). Gel filtration molecular mass standards were obtained from Sigma (Poole, Dorset, UK). SDS–PAGE was carried out by the Laemmli method [19].

2.4 Sequence analysis

Plasmid DNA was prepared following the manufacturer's guidelines (Promega UK Ltd, Southampton, UK). All the necessary buffers were supplied as part of the Wizard Plus SV? minipreps kit. Plasmid DNA was sequenced in both directions on an ABI Prism 377 DNA Sequencer, using an ABI Prism BigDye terminator sequencing ready reaction kit?. DNA sequencing profiles obtained from the DNA sequencer were analysed by ABIView (http://users.cloud9.net/dhk/abiview.html). Protein translation was carried out using The Protein Machine (http://www2.ebi.ac.uk/translate/). Homology searches were carried out using Blast from NCIMB (http://www.ncbi.nlm.nih.gov/blast/blast.cgi). The alignments of DNA and protein sequences were performed by ClustalW analysis (http://www2.ebi.ac.uk/clustalw/).

3 Results

Untransformed CM107 displayed no dipeptidase activity, whereas Pep581 had significant dipeptidase activity against Ala2 (Table 1). The untransformed CM107 possessed some tripeptidase and oligopeptidase activity; this activity was significantly increased in Pep581 (Table 1). Leu p-nitroanilide (Leu-pNA) was the most rapidly hydrolysed of amino acyl-pNA substrates, although a wide specificity was indicated (Table 2). Less activity was detected against dipeptidyl peptidase substrates, except where Ala2-pNA was hydrolysed more rapidly than Ala-pNA, at a rate only slightly lower than Leu-pNA (Table 2). Pep581 was unable to hydrolyse acetylated Ala-pNA but readily hydrolysed Ala-pNA (Table 2), indicating that it was an exopeptidase. Free alanine rather than Ala2 was released, indicating that the enzyme was an amino acyl aminopeptidase rather than a DPP. The effect of different classes of inhibitors on the hydrolysis of Leu-pNA by Pep581 was measured (Table 3). Only 1,10-phenanthroline, a metalloprotease inhibitor, had a significant effect on the breakdown of this aminopeptidase substrate.

View this table:
Table 1

Breakdown of di-, tri- and penta-alanine peptides by clone Pep581 from P. albensis M384 compared with wild-type CM107

Rate of breakdown (nmol/min/mg protein−1)
  • nd = not detectable. Results are the mean of triplicate measurements.

View this table:
Table 2

Activity of clone Pep581 compared with parental CM107 against different amino acyl peptidase and dipeptidyl peptidase substrates

Rate of breakdown (nmol/min/mg protein)
  • nd = not detectable. Results are the mean of triplicate measurements.

View this table:
Table 3

Effect of inhibitors on peptidase activity of clone Pep581

InhibitorClass of inhibitor% of Activity against Leu-pNA
10 mM PMSFSerine protease61
1 mM DCISerine protease68
0.1 mM TLCKTrypsin-like serine protease94
50 μM Pepstatin AAspartyl protease100
10 mM IodoacetateCysteine protease100
50 μM E-64Cysteine protease100
25 mM EDTAMetallo-protease71
10 mM 1,10- phenanthrolineMetallo-protease2
0.1 mM BestatinAmino acyl peptidase71
0.2 mM BenserazideSubstrate analog for Ala-DPP99
0.2 mM Ala2 CMKSubstrate analog for DPP-I73
0.2 mM GPDSubstrate analog for DPP-I57
0.25 mM Diprotin ASubstrate analogue for DPP-IV93
  • Results are the mean of triplicate incubations. The concentration of inhibitor shown is the final concentration in the incubation mixture. 100% activity against Leu-pNA is 20 nmol/min/mg protein. Results are the mean of triplicate measurements. Ala2CMK = Ala2-chlorymethyl ketone, GPD = GlyPhe-diazomethyl ketone, PMSF = phenylmethylsulfonyl fluoride, DCI = 3,4-dichloroisocoumarin, TLCK = tosyl lysyl chloromethyl ketone, E-64 =l-trans-epoxysuccinyl-leucylamide-(4-guanidino)-butane, EDTA = ethylenediaminetetraacetic acid.

Sequence analysis of the plasmid revealed an ORF that showed considerable similarity to a broad specificity dipeptidase, PepD. The sequence has been deposited with GenBank Accession No. AJ867214. The putative pepD gene (Fig. 1) was 1452 bp in length, and encoded a 484-amino acid protein, which had a calculated molecular weight of 53.2 kDa and a theoretical pI of 4.90. A small amount of upstream sequence was obtained, which contained an in-frame stop codon at −4 to −6. Gel filtration indicated a native molecular mass of 115 kDa. SDS–PAGE of crude lysates indicated that the major band had a Mr of 52,000. The upstream region was not determined, although the downstream region contained several stops in every frame. A Kyte and Doolittle plot of the protein (not shown) indicated that the protein was relatively hydrophilic in nature throughout its sequence. The similarity of this putative PepD with other known proteins was determined using BlastP and compared with other known proteins in the databases. The amino acid sequence of PepD from E. coli showed the greatest similarity (47% identity). ClustalW alignment of the amino acid sequence of Pep581 from P. albensis with those of other known PepD amino acid sequences showed many conserved regions (Fig. 2). Most bacterial sequences were very similar to Pep581 from P. albensis.

Figure 1

The nucleotide and derived amino acid sequence of the putative pepD gene from clone Pep581.The start and stop codons are underlined. A putative catalytic metal binding region is highlighted.

Figure 2

ClustalW alignment of PepD amino acid sequences from P. albensis, Haemophilus influenzae, Pasteurella multocida, E. coli, Vibrio cholerae and Borrelia burgdorferi. A putative metal binding domain is highlighted.

4 Discussion

A peptidase-deficient mutant of E. coli was used to screen for peptidase clones derived from P. albensis DNA. E. coli CM107 is unable to hydrolyse dipeptides and synthetic oligopeptides with a blocked C-terminus (C.G. Miller, personal communication). Therefore, no activity against either pNA or MNA amino acyl or peptidyl substrates was observed. However, this strain still possesses a PepT which can hydrolyse tripeptides and also Dcp and OpdA which are carboxypeptidases that can hydrolyse peptides which have not had their C-terminus modified [20]. Thus, background activity against Ala3 and Ala5 was observed in E. coli CM107. The increase in activity with Pep581 was measured by disappearance of the substrate. Thus, it is unlikely that the observed activities against the Ala-oligopeptides could result from, for example, synergistic activity with E. coli enzymes.

The peptidase encoded by the ORF in Pep581 showed strong sequence similarity to the PepD of E. coli and other bacteria. Bacterial PepD enzymes were originally detected as carnosinases, capable of hydrolysing the dipeptide β-Ala-L-His, a naturally occurring peptide found in skeletal muscle, and were found in several different bacteria [21]. Only PepD from E. coli has been characterised fully at the biochemical and genetic level [21,22]. This cytoplasmic enzyme from E. coli is capable of hydrolysing a broad spectrum of dipeptides, providing they have a free N-terminus, but is unable to break down tripeptides [21,22]. In some instances it can also hydrolyse peptides that have had their C-terminus protected with methyl esters or dipeptide amines [21,22]. Metal chelators such as EDTA and 1,10-phenanthroline have a potent inhibitory effect on activity. Pep581 is similar in being a metallo-dipeptidase, but differs in that it hydrolyses longer peptides as an aminopeptidase. The E. coli pepD gene encodes a protein with an Mr of 52 kDa which exists as a dimer in its native form [21]. The calculated Mr of Pep581 was 53.2 kDa, and it was also a dimer. Pep581 displayed a preference for leucine-containing peptides and rapidly hydrolysed the test substrates for PepA, Leu-MNA and Leu-pNA [23]. However, it was unable to hydrolyse LeuVal-MNA, although this was probably due to its difficulty in hydrolysing valine peptide bonds, as demonstrated by its low activity against Val-pNA. This clone was also unable to hydrolyse the DPP-4 substrates GlyPro-pNA, GlyPro-MNA and LysPro-MNA, although it was able to break down Pro-pNA. This would indicate difficulty in cleaving the bond N-terminally of the proline residue. Proline at the N-terminus of peptides renders them resistant to hydrolysis in the rumen [6,24,25].

Bacterial PepDs have as yet to be assigned a NC-IUBMB enzyme classification number, but they have been assigned to family M25 within clan MH of the peptidases [21]. The MH clan of peptidases are all metallopeptidases that do not contain the consensus motif associated with the PepN type aminopeptidases, HEXXH, (where X is any amino acid) [26]. The MH clan also includes a range of other broad specificity peptidases, including the tripeptidase PepT, the dipeptidase PepV, the oligopeptidase PepA and several carboxypeptidases (see Merops database, http://www.merops.co.uk/merops/merops.htm). The exact catalytic site and metal binding residues remain to be determined in the PepD enzyme, but it is interesting to note that a consensus sequence consisting of eight amino acids, similar to that of PepA, (NTDAEGRL) – N(T/V)D(S/T/G)E(E/Q/D)(I/N/E)G – can be identified in all of the bacteria of this family (Fig. 2). Whether this sequence is involved in metal binding in PepD remains to be determined. Nevertheless, because the D and E residues are involved in metal binding in PepA, it would seem logical to assume that they carry out a similar role in PepD.

The capacity of Pep581 to hydrolyse di-, tri- and oligopeptides may indicate a different evolutionary origin to the PepD of E. coli. PepA, PepT and PepD all belong to the same clan of metallopeptidases and may have originated from a similar evolutionary source. Alignment of pepT from E. coli and Pep581 (data not shown) revealed regions of identity, especially over a 40-amino acid long region that has already been speculated to contain the metal binding region [27]. A similar catalytic and metal binding motif is identified in pepN and pepD. The catalytic and metal binding motif has not been determined for pepT, nor has the metal cation necessary for catalysis [27]. However, PepV from Lactobacillus delbrueckii is another broad specificity peptidase, also assigned to clan MH and displays similar biochemical characteristics and inhibitor profiles to PepD [28]. PepV is also capable of hydrolysing tripeptides in addition to dipeptides. Alignment of PepV and PepD from P. albensis revealed significant similarity (data not shown). Thus, although Pep581 from P. albensis is most similar to PepD, it also shares sequence and catalytic similarity to related peptidases.

This is only the second peptidase from a Prevotella sp. to have been cloned, sequenced, and characterised enzymatically. The Mr of the peptidase is consistent with that of the major dipeptidase found in P. albensis[9]. The peptide-hydrolysing activity of Prevotella spp. is an important component in the nutritionally wasteful (for the host animal) breakdown of protein to ammonia in the rumen [2,57]. DPP are highly significant both in Prevotella and in the mixed ruminal community [2,6,8,16]. Amino acyl amino peptidases also occur in the mixed community, and to a lesser extent in Prevotella[8], and metallo-dipeptidases and tripeptidases are significant in both [9,29]. It seems likely, therefore, that the enzyme investigated here may play a role principally in the latter activities.


This work was supported by the Scottish Executive Environment and Rural Affairs Department. The advice of C.G. Miller is gratefully acknowledged.


  1. [1].
  2. [2].
  3. [3].
  4. [4].
  5. [5].
  6. [6].
  7. [7].
  8. [8].
  9. [9].
  10. [10].
  11. [11].
  12. [12].
  13. [13].
  14. [14].
  15. [15].
  16. [16].
  17. [17].
  18. [18].
  19. [19].
  20. [20].
  21. [21].
  22. [22].
  23. [23].
  24. [24].
  25. [25].
  26. [26].
  27. [27].
  28. [28].
  29. [29].
View Abstract