Introduction

Price and collaborators (2006; preceding article ) have recently reported an interesting and exhaustive analysis of his gene orders in a large number of bacterial genomes deposited in the MicrobesOnline database ( http://www.microbesonline.org). This, coupled with a phylogenetic analysis using distance-based methods of the His concatamers, showed that a unified hisGDC(NB)BHAF(IE) operon is present in many lineages. This result has led the authors to hypothesize that this operon is ancient, an assumption that is in partial disagreement with our previous results (Fani et al. 2005). However, our work did not focus on the evolution of his operons in general (i.e., in all completed genomes) but on the construction of the proteobacterial operons. Our main conclusion was that the E. coli-type operon was constructed piecewise during proteobacterial evolution starting from a proteobacterial ancestor whose genome probably contained a set of scattered his genes (or only partially clustered; see below the “his core,” a cluster of four his genes arranged in the order hisBHAF [Fani et al. 1995]. This does not per se imply that the Last Universal Common Ancestor possessed scattered his genes.

The analysis of gene order is of great importance to understand the evolution of operon structures; however, it should be integrated with other information, such as gene regulative controls and the presence of fused genes. In this sense, histidine biosynthesis represents an excellent study model since it has been investigated for over 40 years and a wealth of experimental and bioinformatics results is available. Therefore, we did not limit our study to the analysis of strings of sequences, but also considered other important aspects of his genes (e.g., presence/absence of given gene fusions, type of transcriptional regulation).

If we look in detail at the results of Price et al. (2006), we see that they support that his core is present in the proteobacterial ancestor, but not that the E. coli-type operon is ancient and common in other bacterial branches, because it has peculiar features.

The term unified operon introduced by Price et al. seems to indicate very similar structures in different organisms, but his operons may have very peculiar features in different taxonomic lineages, and these have to be taken into account when deriving the evolutionary history of these biosynthetic operons.

Gene Order

Price et al. propose that the unified E. coli type his operon is present in the Euryarchaeota; this is not true, since these archaeal operons lack the hisNB fusion and include hisZ, a gene which is not present in the E. coli genome (see below). In addition, there are hints concerning the Pyrococcus furiosus/Thermococcus kodakarensis and Ferroplasma acidarmanus/Picrophilus torridus operons suggesting that they were acquired by lateral gene transfer from a bacterial donor (Brilli and Fant, manuscript in preparation). Actually, these operons show complete synteny conservation with some bacterial operons, parallel with an evident divergence from all the other Archaea (Fig. 1). This is in agreement with data from Huynen and Bork (1998), who stated that “gene order is rarely conserved in evolution,” hence “the presence in two distant evolutionary branches of the same order of genes, combined with the absence of this gene order in other more closely related branches, can point to horizontal transfer.” Moreover, the his operon of these Euryarchaeota includes both a bifunctional hislE gene and hisZ, the latter coding for a regulatory protein. Both of them are present in these four species (of two taxonomic groups) but never in other archaeal species, while both are common in Bacteria (Sissler et al. 1999; Vega et al. 2005). Finally, all the closest relatives of the above-mentioned archaeal species lack the entire set of histidine biosynthetic genes. The crenarchaeal Sulfolobales-like operon (hisCGABFDEHI) shows a gene order that is completely different from the bacterial type and it lacks the his core. The gene order of partial his clusters in other Archaea often resembles the Sulfolobales operon gene order but not bacterial gene orders (e.g., hisAG in Methanopyrus kandleri, hisGAB in Methanosarcina mazei, hisAB in Archaeoglobus fulgidus; these genes have never been found in these orders on bacterial chromosomes). These features, along with phylogenetic results, seem to indicate that the resident archaeal his operon is represented by the Sulfolobus type, while the core containing operons appeared in Bacteria and were later transferred to some Archaea.

Fig. 1.
figure 1

Organization of his genes in some representatives of Archaea and Bacteria. The E. coli-like hisG gene is in black, the hisZG and hisZ genes are shown in dark and light gray, respectively. A gray bar indicates the genes forming the core of histidine biosynthesis.

Full size image

Gene order differences in operons also occur in Bacteria, but a constant theme across different taxonomic groups is the presence of the his core, hence a construction of the his core outside, and before, proteobacterial divergence is possible (Brilli and Fani 2004; Fani et al. 2005). In our paper (Fani et al. 2005) we did not completely rule out this possibility and the new available genomes support this view. Nonetheless, his operons contain more genes than those in the core and their gene order is not as constant as the analysis by Price et al. (2006) seems to indicate: e.g., in some Firmicutes the hisC gene is located outside the operon and belongs to a different transcriptional unit (e.g., Bacillus subtilis and Zymomonas mobilis [Gu et al. 1995], containing trp, aro, and tyr genes, which are involved in other biosynthetic pathways HisC is involved in, in addition to histidine biosynthesis; on the contrary, in other Firmicutes, the hisC gene is the first of the operon and it is not involved in other pathways (e.g., Lactococcus lactis [Delorme et al. 1992].

Gene Fusions

E. coli-like operons contain the hisNB gene fusion, which is peculiar to γ-proteobacteria, except for documented cases of lateral gene transfer, such as Campylobacter jejuni (Brilli and Fani 2004). Thus, it is quite unlikely that this fusion was present in the last common ancestor of Bacteria (or Proteobacteria) and then lost (often independently; see the α/β/δ/ε-proteobacterial branches) in all but one (the γ-proteobacteria) taxonomic branch (see Brilli and Fani 2004).

The HOL-Pase Coding Gene

The hisN gene codes for a histidinol-phosphate phosphatase (HOL-Pase), and this activity has been ascribed to at least three different (and phylogenetically unrelated) types of enzymes, depending on the (group of) organisms considered: in at least some Firmicutes, the budding and fission yeasts, this enzyme belongs to the PHP (polymerases and histidinol phosphatases) family (le Coq et al. 1999; German and Hu 1969; Millay and Huston 1973). ln the γ-proteobacteria (and probably in other proteobacterial branches) it is part of the DDDD superfamily (Brilli and Fani 2004 and references therein). Moreover, in Neurospora crassa this activity is probably carried out by an aspecific alkaline phosphatase (Morales et al. 2000). ln most organisms the gene coding for this enzyme is completely unknown and often the sequence corresponding to known types of HOL-Pases cannot be retrieved in complete genomes. As previously suggested (Brilli and Fani 2004), this gene was very likely recruited in a later evolutionary stage of histidine biosynthesis evolution. Different types of phosphatases are devoted to histidinol dephosphorylation in different groups of organisms, and the position of the corresponding genes in his operons is highly variable, even between closely related organisms (i.e., Gram positives).

Feedback Inhibition

In E. coli and the other γ-proteobacteria the HisG enzyme is able to autoregulate its activity by binding histidine when the intracellular amino acid concentration exceeds a threshold level. A few years ago a different feedback inhibitory mechanism of histidine biosynthesis was disclosed in L. lactis (Sissler et al. 1999) and then found in an increasing number of microorganisms. This mechanism is based on the interaction between HisZ and a shorter version of the HisG protein, which is per se unable to bind histidine, a function performed by HisZ, which then turns off HisG. Thus, HisZ regulates HisG and it is not present in γ-proteobacteria (Kleeman and Parsons 1977; Boveeé et al. 2002; Vega et al. 2005; Lohkamp et al. 2004; Fani et al. manuscript in preparation), representing a remarkable difference in the feedback regulatory mechanism of histidine biosynthesis that impacted on operon evolution too: if present, the hisZ and hisG genes are very often contiguous (and overlapping) on the genome. If clustered in the operon, the two genes occupy the first two positions.

Transcriptional Regulation

The ways in which his operon transcription is regulated can also vary among different organisms: several Firmicutes and γ-Proteobacteria possess an attenuator sequence upstream of his operons. However, the mechanisms are different in the two groups: the E. coll attenuator region contains seven his codons, which are directly involved in the attenuation mechanism (Alifano et al. 1997 and references therein), while the Lactococcus lactis his operon promoter lacks significant pairings with the anticodon of the tRNAhis (Delorme et al. 1999) and histidine codons are not retrieved in its sequence.

All these differences indicate that also starting from partial clusters in the proteobacterial (or the LC) ancestor, the evolutionary history of the E. coli-like operons appears to be unique, as are many of its peculiar features.

Phylogenetic Analysis

Price et al. performed a phylogeny of the concatenated His proteins using a distance method. This approach has the main disadvantage, with respect to maximum likelihood (ML), that concatenated sequences are transformed into a single distance matrix, losing the information on each protein’s effective contribution to the phylogeny. Moreover, the information in alignment columns containing gaps is not used. It is noteworthy that distance methods are more susceptible to long-branch attraction (Felsenstein 2004) than ML; this may happen with diverging amino acid sequences. Several studies show that ML methods generally outperform distance methods over a broad range of realistic conditions (Kuhner and Felsenstein 1994; Gaut and Lewis 1995; Huelsenbeck 1995).

Finally, we note that species such as Aquifex, the cyanobacteria and the T/D group have scattered genes. If the complete operon is really ancient, we have to postulate several independent partial or total destructions of the operon in these species and differential maintenance in those which differentiated in a later stage.

Conclusions

The scenario of a proteobacterial ancestor harboring a his core (hisBHAF), and of different genes being joined to this aggregation nucleus at different times, seems to be better supported by sequence and other functional data than other hypotheses.

The presence of a completely formed E. coli-like his operon very early in evolution does not appear to be supported by the comparative analysis of genomes and by the analysis of gene structure, organization, and regulation. This indicates that if this view is correct, then a great number of independent rearrangements would be necessary to explain the extant situation.

We cannot a priori exclude—and it was beyond the scope of the paper by Fani et al. (2005)—the possibility that a partial clustering of his genes occurred in the early stages of bacterial evolution. However, the E. coli-type operons containing all his genes in the order hisGDC(NB)HAF(IE) emerged later during evolution, as did the fine regulatory mechanisms which control transcription today in many species and the gene fusions found in the γ branch.

By this we can explain the present-day differences as different evolutionary outcomes which all started from a common set of histidine biosynthetic genes, but which happened in independent ways in the different lineages, in conformity with the different organisms’ needs, which are directly related to the ecological niche the organism itself lives in.