Vehicular Edge Computing (VEC) offers a promising paradigm for low-latency and high-bandwidth services by extending cloud capabilities to the network edge. A critical challenge in VEC is the dynamic placement of microservices on Edge Servers (ESs) due to the high mobility of vehicles. This paper proposes a Deep reinforcement learning (DRL) approach, specifically using Proximal Policy Optimization (PPO), to address the microservice placement problem in a VEC environment. Our objective is to minimize service latency and resource consumption while adapting to dynamic traffic conditions. Real-world vehicle traces from Luxembourg City are obtained via the SUMO simulator to create a realistic environment for training and evaluation. Experimental results demonstrate that the proposed PPO-based framework achieves superior latency resource trade-offs, adapts effectively to mobility variations, and outperforms traditional heuristic and DRL baselines.