**5. Conclusions**

The monophyly of eukaryotic and cyanobacterial GOX-like proteins points to a cyanobacterial origin of all eukaryotic GOX-like proteins, including the photorespiratory GOX of phototrophic eukaryotes. Biochemically, these proteins operate as GOX rather than LOX in the Archaeplastida. The slight preference for glycolate of the synthetic ancestral N3-GOX suggests that glycolate oxidation could have been the ancestral function, and during the course of evolution, this protein has changed its preference to glycolate or l-lactate and other related substrates according to the requirements of the specific host.

The likely scenario for the evolution of GOX among eukaryotes and particularly of the photorespiratory GOX of Archaeplastida is summarized in Figure 6. A bifunctional GOX-like protein (represented by N3-GOX) was transferred from an ancient cyanobacterium via HGT into eukaryotes before the split of animal, fungal, and plant lineages occurred. Most extant cyanobacteria lost this enzyme, and use a glycolate dehydrogenase for photorespiratory glycolate oxidation, whereas N2-fixing cyanobacteria optimized the ancestral GOX-like protein to operate as LOX, which helps consume oxygen, protecting nitrogenase [30]. Among the Metazoa, the ancestral cyanobacterial protein evolved into GOX-like enzymes with varying substrate specificities after gene duplication and biochemical specialization, as suggested by Esser et al. [32]. After the engulfment of an ancient cyanobacterium as a plastid ancestor, possibly two gene copies for GOX-like proteins co-existed in the proto-alga, but only one of these copies was retained. Among the Archaeplastida, the ancestral GOX-like protein early on evolved to the main photorespiratory glycolate oxidizing enzyme and became localized in peroxisomes. The evolution of peroxisomes in eukaryotes is still a matter of discussion, but it has been shown that the proteome of peroxisomes is variable, pointing to an evolutionary optimization of peroxisome functions in time by protein acquisitions and losses [74]. The increasing photorespiratory flux due to the increasing oxygen concentration in the environment was matched by increasing the Vmax of glycolate oxidation.

**Figure 6.** Hypothetical evolutionary scenario of GOX-like proteins of the 2-hydroxy-acid oxidase family among Eukaryotes. A bifunctional GOX-like protein and a glycolate dehydrogenase (GlcDH) existed in ancient cyanobacteria. In most extant cyanobacteria the GOX-like gene was lost, and they use GlcDH for photorespiratory glycolate oxidation. The cyanobacterial gene of the GOX-like protein was initially transferred to the eukaryotic genome via horizontal (endosymbiotic) gene transfer (HGT/EGT). After the engulfment of an ancient cyanobacterium as a plastid ancestor, probably two gene copies for GOX-like proteins existed. Subsequently, one of these copies was lost during plastid establishment. The early evolution of the glycolate specificity of the GOX-like protein most likely took place in the proto-alga before the split-off of Archaeplastida lineages. Only among cyanobacteria and chlorophytes, did GOX-like proteins evolve into l-lactate oxidase (LOX).

Hence, the early acquisition of a peroxisomal GOX-like protein by the hypothetical host cell [75–77] can be regarded as further exaptation to integrate the photosynthetic plastid ancestor with an already active host glycolate metabolism in the oxygen-containing atmosphere [22].

**Supplementary Materials:** The following are available online at http://www.mdpi.com/2223-7747/9/1/106/s1, Supplementary material 1: Figure S1: GOX-like proteins from eukaryotes and prokaryotes with different substrate preferences share a common tertiary and quaternary structure; Figure S2: Sequence of the cDNA and protein of the GOX from *Spirogyra pratensis* used in this study; Figure S3: Corrected sequence of the Cp-GOX coding gene in *Cyanophora paradoxa*; Figure S4: Collapsed phylogenetic tree of GOX-like proteins based on 111 amino acid sequences from all organismic groups in the tree of life; Figure S5: Alternative rooting of phylogenetic trees of GOX- and LOX-like sequences; Figure S6: Sequence of the protein and the codon-optimized DNA of the ancestral GOX-like protein (N3-GOX) used in this study; Figure S7: Schematic display of the work to obtain the correct complete *cp*-*goxc* gene from *Cyanophora paradoxa;* Figure S8: Purification of recombinant GOX proteins from *Spirogyra* (Sp-GOX) and *Cyanophora* (Cp-GOXc) and the ancestral GOX protein (N3-GOX); Table S1: Species names and accession numbers for phylogeny; Table S2: Catalytic efficiency of recombinant GOX and GOX-like proteins; Table S3: Sequences of the used primer sequences Supplementary Material 2: Alignment of GOX-like proteins, accession numbers: Sp-GOX: AVP27295.1 and Cp-GOX: AVP27296.

**Author Contributions:** Conceptualization, R.K. M.H., and H.B.; methodology and investigation, R.K. and F.F.; validation and interpretation, R.K. and M.H.; formal analysis, R.K., M.H., C.D., and A.P.M.W.; resources, A.P.M.W., H.B., and C.D.; writing (original draft), R.K. and M.H.; writing (review and editing), all co-authors; supervision, M.H., A.P.M.W., and H.B. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research was funded by the DFG (Deutsche Forschungsgemeinschaft), in the frame of the Forschergruppe FOR 1186-Promics.

**Acknowledgments:** We thank Endymion D. Cooper (University of Maryland, USA) for his help obtaining the Sp-GOX cDNA sequence. We appreciate the technical assistance of Manja Henneberg and the discussions with Lukas Krebes.

**Conflicts of Interest:** The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results.
